applied sciences

Article Security Assessment of Urban Areas through a GIS-Based Analysis of Lighting Data Generated by IoT Sensors

Lavinia Chiara Tagliabue 1,* , Fulvio Re Cecconi 2 , Nicola Moretti 2 , Stefano Rinaldi 3 , Paolo Bellagente 3 and Angelo Luigi Camillo Ciribini 1 1 Department of Civil, Environmental, Architectural Engineering and Mathematics, University of Brescia, 25123 Brescia, Italy; [email protected] 2 Department of Architecture, Build Environment and Construction Engineering, Politecnico di Milano, 20133 Milan, Italy; [email protected] (F.R.C.); [email protected] (N.M.) 3 Department of Information Engineering, University of Brescia, 25123 Brescia, Italy; [email protected] (S.R.); [email protected] (P.B.) * Correspondence: [email protected]; Tel.: +39-02-2399-6018



Received: 27 January 2020; Accepted: 12 March 2020; Published: 23 March 2020


Abstract: The current perspective about urban development expects 70% of energy consumption will be concentrated in the cities in 2050. In addition, a growing density of people in the urban context leads to the need for increased security and safety for citizens, which imply a better lighting infrastructure. Smart solutions are required to optimize the corresponding energy effort. In developing countries, the cities’ lighting is limited and the lighting world map is strongly significant about the urban density of the different areas. Nevertheless, in territories where the illumination level is particularly high, such as urban contexts, the conditions are not homogenous at the microscale level and the perceived security is affected by artificial urban lighting. As an example, 27.2% of the families living in the city of Milan, ombardy Region, Italy, consider critical the conditions of lighting in the city during the night, although the region has diffused infrastructure. The paper aims to provide a local illuminance geographic information system (GIS) mapping at the neighborhood level that can be extended to the urban context. Such an approach could unveil the need to increase lighting to enhance the perceived safety and security for the citizens and promote a higher quality of life in the smart city. Lighting mapping can be matched with car accident mapping of cities and could be extended to perceived security among pedestrians in urban roads and green areas, also related to degradation signs of the built environment. In addition, such an approach could open new scenarios to the adaptive street lighting control used to reduce the energy consumption in a smart city: the perceived security of an area could be used as an additional index to be considered during the modulation of the level of the luminosity of street lighting. An example of a measurement set-up is described and tested at the district level to define how to implement an extensive monitoring campaign based on an extended research schema.

Keywords: GIS; open data; security mapping; lighting mapping; perceived security; user centered design; mobile sensing; urban sensing; IoT sensors

1. Introduction Urban centers are becoming increasingly attractive for people in developing and developed countries and the transitions of people toward the cities has rapidly grown in the past 40 years. In 1950, in general about 2/3 of the world population lived in rural settlements and 1/3 was living in urban areas to exploit working conditions and opportunities. The share of the global population living in urban

Appl. Sci. 2020, 10, 2174; doi:10.3390/app10062174 www.mdpi.com/journal/applsci Appl. Sci. 2020, 10, 2174 2 of 30 areas has increased from one third in 1960 to 55% (4.1 billion people) in 2018 [1]. The world’s urban population is now increasing by 60 million persons per year, about three times the increase registered in the rural population [2]. Growing urbanization is a consequence of the new births in urban zones and of continued migration of people coming from the rural surroundings. These movements are also enhancing the sprawl of urban spaces as before rural peri-urban settlements were combined into nearby cities and secondary cities, connected by commerce activities developed in larger urban centers [3]. The process of urbanization is expected to strongly endure in the next century. By 2030, it is projected that nearly 5 billion of the world’s 8.1 billion people will live in urban developments, which means 61% of the population. The less-developed regions will be more than 57% urban. Latin America and the Caribbean will actually have a greater percentage of inhabitants living in cities than in Europe [4] because of the disparity of conditions and services that people can find in the city centers. In the past, the movement from rural areas to the cities could guarantee better living conditions, however there is scientific evidence that now this assumption is not always valid. There is a high disparity in living conditions between downtown and the surroundings in US [5], European and Asian cities [6]. Worldwide, the number of metropolises with 10 million or more inhabitants is increasing rapidly, configuring new ‘megacities’ located in the less-developed areas of the world. In 1960, only New York and Tokyo had more than 10 million people, by 1999 the number of cities with more than 10 million people around the world was 17, with 13 located in developing regions [7]. Currently, urban centers occupy less than 5% of the world’s mainland, using over two-thirds of the total energy, and they are responsible for over 70% of CO2 emissions, according to the C40 Cities Climate Leadership Group [8]. In 2050, the outlook is that 6 billion people will live in urban zones, worldwide. Thus, innovation in urban infrastructure and technology is essential when addressing sustainable and low-impact solutions to support anthropic activities and urban growth and to reduce the cities’ impact. As an example, by 2030, greenhouse gas (GHG) emissions could be reduced by up to 1.5 billion metric tons CO2eq annually, mainly through a massive transformative change in the transport systems in the world’s 724 largest cities [9]. The availability of a pervasive communication infrastructure, wireless [10] or cable [11], in the city of the future is a prerequisite to enable services and solutions to improve sustainability in the use of the energy resources [12,13]. How to manage this incomparable urban expansion in following years is likely to decide the outcome of the sustainability endeavors that are spreading through different sectors [14]. The global trends of urbanization in the first decades of the 21st century are meaningfully transformed from the preceding experiences in terms of urban change. For instance, urbanization is taking place at lower levels of economic development and the majority of future urban population increase will take place in developing countries where the aggregation of people in the urban centers often occurs in informal settlements [15]. These kinds of urban forms of organization are taken in consideration to promote upgrading and regeneration of social conditions and health improvements, considering that in some cases the term slums is used to define such casual and degraded environment where security issues are one of the first matters to consider [16]. The expansion of urban areas is on average twice as fast as urban population with significant concerns for GHG emissions and climate change effect, as mentioned before [17] but also for other kinds of issues and pollution problems related to urban aggregation (i.e., air, soil, water, light pollution). The present research focuses on outdoor artificial lighting, a variable that is closely associated with urbanization because it has interdependencies and outcome at different levels: energy issues, social impacts, use of the city, and psychological perceptions of security. Outdoor artificial lighting is strongly connected with urbanization and anthropic development of territory [18]. The use of lighting in the urban context is extensively adopted in large urban centers and dark areas are experiences as disturbing by citizens while in rural areas people are used to moving in darker areas. Additionally, lighting is strongly related to the assessment of safety related to traffic management [19] and people security in the cities. In particular, the lighting system has a strong Appl. Sci. 2020, 10, 2174 3 of 30 impact on vehicle accidents, as clearly demonstrated in [20–22]. The correlation between the quality of the light system and the number of car accidents is well investigated in the literature [23–25]. A low level of or inadequate street lighting is among the first cause of road accident because it has a strong impact on the attention of the drivers. For this reason, the best practice for the design of a new streetlight infrastructure takes into consideration several parameters, including the typology of the road (main or subsidiary), the typical users of the road and the presence of pedestrian and bicycle [26]. The importance of monitoring the quality of light for preventing road accidents is demonstrated in previous research work [27]. The research shows that financial savings from reduced crimes through improved lighting significantly overcame the financial costs of the upgraded street lighting. A meta-analysis found that improved lighting led to reductions in crime of 20% in experimental areas compared with control areas. However, night-time crimes did not decrease more than daytime crimes, and a theory focusing on the role of street lighting in increasing community pride seems more plausible than a theory focusing on amplified surveillance, supporting the regenerative role of lighting in the urban areas. Also ecological systems are directly affected by lighting pollution [28] whilst indirect effects are registered on economics [29] and carbon emissions [30] related to urban aggregation and population growth. The weight of lighting in the energy consumption of the diverse countries varies even if it can reach 40% [31,32] of the total energy consumption of the city with an environmental impact ranging from 13% to 37% of GHG emissions for the municipal sector [33]. Given the wide-ranging sustainability inferences of the rise of artificial lighting, research programs are evolving that inspect these phenomena from a range of disciplinary viewpoints. Nonetheless, strategies and policies for the artificial lighting management [34] are less complete than might be predictable [35]. The absence of high-resolution mapping of artificial lighting is progressively documented as an important barrier to related research and management. In contrast, the remote sensing of nocturnal lighting offers an accurate, economical, and straightforward way to map the global distribution and density of developed areas [36]. Datasets exist globally at a coarse spatial (~3 km) and spectral resolution, allowing us to understand the broad variations in urban lighting [37] but sub-city patterning cannot be explored effectively [38,39]; however, important correlations can be found between population, road density and lengths, and people’s incomes. Numerous satellite images are available from the International Space Station with a spatial resolution of up to 60 m. Although these images are starting to be used to detect demographic patterns within urban areas, and in principle, such a mosaic image could also be used to automatically identify and classify the different lights of the city, in practice, individual lamps still cannot be identified [40] and there are some difficulties associated with imaging light pollution: large dynamic range, rapid motion of the sensors, need to sample a large area in a short period of time (requiring a large field of view, FOV). Finer spatial resolution data do exist, but typically have a limited spatial extent [41] along a landscape modification gradient where the three following phenomena have been recognized [42]: (1) patch density increases exponentially; (2) the regularity of patch shape increases; (3) landscape connectivity decreases [43,44]. These patterns have been constantly supported for urbanizing landscapes throughout the world, however this hinders the development of a robust evidence base to sustain urban lighting policies, as cities can be highly heterogeneous even at fine spatial scales [44]. For example, little is known about how lighting varies with urban land use [45] or alongside built density gradients. Upgraded baseline urban lighting maps are similarly required in order to apply the outcomes of published lighting research [46], to implement existing planning guidance on urban lighting zones, [47], to implement planning agreements and legislation standards related to lighting nuisance [48] and to monitor changes over time. Consequently, we need confident lighting datasets at the city scale; and with a spatial and spectral resolution that should be sufficient to advance lighting research and the development of planning and governance of urban lighting. The aim of the current research work is to remedy to the current lack of lighting data in urban areas with a proper spatial resolution. Lighting data are missing because measurement campaigns are Appl. Sci. 2020, 10, 2174 4 of 30 generally expensive, and this is because they involve several technicians moving around different area of cities (and in particular the most unsafe parts of them) during the night with proper measurement devices and the logging of this data in databases. The proposed approach tries to automate this process thanks to the use of a geographic information system (GIS)-based informative system, integrated with lighting data generated by mobile Internet of Things (IoT) sensors The data generated by the IoT sensors can be automatically inserted in a GIS system. Such an approach allows data to be correlated about lighting quality with respect to the perceived level of security and with quantitative data about car accidents and security events. The integration of IoT sensors with the remote GIS system is performed using state of the art IoT protocols [49]. At the present, a measurement prototype based on the Arduino microcontroller has been designed. The IoT sensor is mounted on a mobile mechanical structure which allows easy movement of the sensors in urban areas. The IoT sensor is equipped with off-the-shelf light sensors based on a photodetector. The use of an appropriate sensor calibrated using a V-lambda filter and using cosine correction response input optics, as required by the international norm CIE-194:2011 [50], will provide more accurate light measurement but is out of the scope of the current research work. In fact, as mentioned above, the main contribution of the current research lies in the definition of a GIS-based methodology to analyze and correlate data, more than in the proposal of a new measurement of light sensors.

2. The Lighting System in Urban Area

2.1. Urban Lighting Typologies and Drawbacks The extensive use of outdoor lighting is quite recent, tracing its origins to the electric light bulb commercialized by Thomas Edison in the early 1880s. The artificial nocturnal lighting enables a remote sensing of atrophic activities with a level of detail depending on some cultural variations in the quantity and quality of lighting in different countries. It is possible to underline a notable level of correspondence in lighting technology and lighting levels around the world. The key factor affecting the quantity of lighting is wealth, which means that regions with high per capita income install much more lighting than those with a lower per capita income. Even within wealthy regions, however, lighting technology (i.e., lamps and lighting fixtures) is increasingly changing as growing pressure is promoted to shrink night-time sky brightness and save energy. In particular, the huge impact of lighting systems in term of energy consumption required the definition of proper performance indicators to identify the optimal energy/lighting tradeoff [51]. The more environmentally friendly lamps are the low-pressure sodium models, followed by high-pressure sodium. The most polluting are the lamps with a strong blue emission, like metal halide and white light-emitting diodes (LEDs) [44]. The shifting from the now broadly used sodium lamps to white lamps (MH and LEDs) would produce an increase of pollution in the scotopic and melatonin suppression bands of more than five times the present levels, supposing the same photopic installed flux. This upsurge will aggravate known and likely unknown effects of light pollution on human health, environment and on visual perception. The suppression of melatonin could increase risk of breast cancer through disruption of circadian cycles and studies on LAN (light-at-night) theory are ongoing [52] and solutions to mitigate the issues are crucial to preserve human health. Lighting is also connected with contemporary culture and amenity concepts notwithstanding light pollution has been described as “one of the most rapidly increasing alterations to the natural environment;” an issue whereby “mankind is proceeding to envelop itself in a luminous fog” [53,54].

2.2. Urban Lighting as Development Indicator Researches on this topic showed as remotely sensed lighting data can be assumed to predict both urban extension [55] considering borders and occupied territory [56] encouraging agreement between lighted areas and various definitions of urban extent [57]. For example, conurbations greater than Appl. Sci. 2020, 10, 2174 5 of 30

80 km of diameter account for <1% of the analyzed settlements but describe about half the total lighted area worldwide. The area/perimeter distributions designate that urban areas become progressively dendritic while growing. Studies about lighted areas with built areas estimated from Landsat imagery in a sample of 17 cities showed that lighted areas are regularly larger than even maximum estimated built areas for almost all cities in every light dataset. Thresholds >90% can often reunite lighted area with built area in the datasets dating from 1994/1995 nevertheless there is not one level that fits the Appl. Sci. 2020, 10, x FOR PEER REVIEW 5 of 31 majority of the sample cities. Outdoorestimated lighting built areas can alsofor almost provide all cities an in estimation every light dataset. of the Thresholds population >90% distribution, can often reunite when rough spatial scaleslighted are area accepted; with built the area Defense in the datasets Meteorological dating from Satellite 1994/1995 Program’s nevertheless Operational there is not one Linescan level System that fits the majority of the sample cities. (DMSP-OLS)-basedOutdoor research lighting providedcan also provide an approximated an estimation of assessment the population of thedistribution, world population when rough that could demonstratespatial a 6% scales of discrepancy are accepted; withthe Defense the correct Meteorological value, using Satellite the Program’s night-time Operational satellite Linescan imagery and the rate of urbanizationSystem (DMSP-OLS)-based of the different research nations provided [58]. an approximated assessment of the world population Withthat increased could demonstrate urbanization, a 6% theof discrepancy spatial coverage with the andcorrect intensity value, using of artificialthe night-time light satellite pollution seem imagery and the rate of urbanization of the different nations 58. to be growingWith [21 ,increased59]; furthermore, urbanization, the the spatial night coverage sky spectrum and intensity is correspondingly of artificial light pollution shifting seem due to the replacementto be of growing urban lightingError! Reference infrastructure source not [60 ].found. Lighting; furthermore, enables the citizens night sky to experiencespectrum is the built environmentcorrespondingly and research shifting into thedue e toffect the onreplac urbanement lighting of urban on lighting behavior infrastructure are trying Error! to unveil Reference the outcomes related to environmentalsource not found..psychology Lighting enables and citizens social to interaction experience the [61 built]. environment and research into the effect on urban lighting on behavior are trying to unveil the outcomes related to environmental Lightingpsychology has strong and social cultural interaction link Error! to ideas Reference of modernity source not found. and safety. whilst developing land-use strategies to containLighting has lighting strong cult pollutionural link to can ideas lead of modernity to energy and andsafety economicwhilst developing saving, land-use mitigation of climate changestrategies as to well contain as preservation lighting pollution of can biodiversity lead to energy and and ecosystem economic saving, services mitigation dependent of climate on natural darkness [change62] (Figure as well1 as). preservation of biodiversity and ecosystem services dependent on natural darkness Error! Reference source not found. (Figure 1). Thus a multi-layeredThus a multi-layered evaluation evaluation is needed.is needed. TheThe artificial artificial lighting lighting is a demonstration is a demonstration of progress, of progress, giving peoplegiving wider people optionswider options about about where, where, when when an andd how how long long their their activities activities can be performed. can be performed. Nevertheless,Nevertheless, lighting lighting has direct has direct effects effects on health on health [63 Error!] and Reference sector studies source demonstratenot found. and that sector the amount of pollutionstudies is strongly demonstrate dependent that the amount on the of spectralpollution is characteristics strongly dependent of on the the lamps spectral [64 characteristics]. of the lamps [64].

Figure 1. Night view of Europe where lighting depicts the urban areas and main connections 65. Figure 1. Night view of Europe where lighting depicts the urban areas and main connections [44,65]. 2.3. Urban Lighting Renovation 2.3. Urban Lighting Renovation The increasing urbanization has led to the request of more and better data at the urban level. The increasingSome examples urbanization of lighting hasrenovation, led to thesuch request as 66, under of more European and better project data funding at the framework, urban level. Some examplesshowed of lighting high level renovation, of possible such improvement as [66], through under local European policies to project reduce fundingthe installed framework, power and showed high level ofthepossible number of improvement lamps achieving through consistent local cost policies (averageto value reduce 60%) the and installed energy savings power (average and the number value 54%) results of the light renovation (Error! Reference source not found.). of lamps achieving consistent cost (average value 60%) and energy savings (average value 54%) results of the light renovation (Figure2). The environmental impact reduction ranges depending on the extension of the renovation intervention (from 4.7 up to 138 kgCO2eq). The number of lamps shows an average reduction of 25% with cases where no reduction is registered up to 72%. It is worth noting that the luminaries points undergone to an average reduction of 0.5% with a maximum value of 6.9% (Figure3). Recent research Appl. Sci. 2020, 10, x FOR PEER REVIEW 6 of 31

Appl. Sci. 2020, 10, 2174 6 of 30

workAppl. Sci. [67 2020] suggests, 10, x FOR alternative PEER REVIEW refurbishment design based on energy prices, able to optimize6 of the 31 investment costs.

(a)

(a)

(b)

Figure 2. (a) Percentage of reduction of installed power for artificial outdoor lighting, annual consumption reduction and annual electricity cost reduction in the European renovation project Streetlight EPC (Energy Performance Contracting) 66 in Austria; (b) energy saving and electricity cost saving in the same cities of the mentioned project.

The environmental impact reduction ranges depending on the extension of the renovation intervention (from 4.7 up to 138 kgCO2eq). The number(b) of lamps shows an average reduction of 25% with cases where no reduction is registered up to 72%. It is worth noting that the luminaries points FigureFigure 2.2. ((aa)) Percentage Percentage of of reduction of installedinstalled powerpower forfor artificialartificial outdooroutdoor lighting,lighting, annualannual undergone to an average reduction of 0.5% with a maximum value of 6.9% (Error! Reference source consumptionconsumption reductionreduction andand annualannual electricityelectricity costcost reductionreduction inin thethe EuropeanEuropean renovationrenovation projectproject not found.). Recent research work 67 suggests alternative refurbishment design based on energy StreetlightStreetlight EPCEPC (Energy(Energy Performance Performance Contracting) Contracting) [66 66] in in Austria; Austria; ( b) energy energy saving and electricity costcost prices, able to optimize the investment costs. savingsaving inin thethe samesame citiescities ofof thethe mentionedmentioned project.project.

The environmental impact reduction ranges depending on the extension of the renovation intervention (from 4.7 up to 138 kgCO2eq). The number of lamps shows an average reduction of 25% with cases where no reduction is registered up to 72%. It is worth noting that the luminaries points undergone to an average reduction of 0.5% with a maximum value of 6.9% (Error! Reference source not found.). Recent research work 67 suggests alternative refurbishment design based on energy prices, able to optimize the investment costs.

(a)

Figure 3. Cont.

(a) Appl.Appl. Sci. Sci. 20202020,, 1010,, x 2174x FORFOR PEERPEER REVIEWREVIEW 77 of of 31 3031

((bb))

FigureFigure 3.3. ((aa)) (Global (Global (Global Warming WarmingWarming Potential PotentialPotential (GWP) (GWP)(GWP) emission emissionemission reduction reductionreduction in inin tons tonstons of ofof CO COCO222 forforfor thethethe AustrianAustrian projectsprojects inin thethe EuropeanEuropeanEuropean renovation renovatirenovationon project projectproject Streetlight StreetlightStreetlight EPC EPCEPC [ 6666;66;]; (( (bbb))) numbernumber number ofof lampslamps beforebefore and and after after thethe renovation renovation andand luminairesluminaires pointspoints ofofof lightlight beforebefore ininin thethe samesame cities.cities. The The luminaires luminaires points points after after the the renovationrenovation areare thethe samesame valuevaluevalue reportedreportreporteded forforfor thethethe newnewnew numbernumbernumber ofofof lamps.lamps.lamps.

3. The Proposed Solution 3.3. TheThe ProposedProposed SolutionSolution As well-known from the literature [68,69], GIS systems are widely used to store and to correlate AsAs well-knownwell-known fromfrom thethe literatureliterature 69,69, GISGIS systemssystems areare widelywidely usedused toto storestore andand toto correlatecorrelate information regarding several aspect of urban infrastructure, including the state of the pavements informationinformation regardingregarding severalseveral aspectaspect ofof urbanurban infrastructure,infrastructure, includingincluding thethe statestate ofof thethe pavementspavements alongside the roads, the electrical grid and much more information. GIS can be used not only to alongsidealongside thethe roads,roads, thethe electricalelectrical gridgrid andand muchmuch moremore information.information. GISGIS cancan bebe usedused notnot onlyonly toto visualizevisualize visualize the data in thematic maps but, by using properly queries, it is possible to correlate spatially thethe datadata inin thematicthematic mapsmaps but,but, byby usingusing properlyproperly queriequeries,s, itit isis possiblepossible toto correlcorrelateate spatiallyspatially andand temporallytemporally and temporally the data about “road accidents” with “luminosity data”, especially if the luminosity thethe datadata aboutabout “road“road accidents”accidents” withwith “luminosity“luminosity data”,data”, especiallyespecially ifif ththee luminosityluminosity datadata areare periodicallyperiodically data are periodically updated. updated.updated. A method for the collection, processing and representation of the light intensity data through the AA methodmethod forfor thethe collection,collection, processingprocessing andand reprrepresentationesentation ofof thethe lightlight intensityintensity datadata throughthrough use of IoT sensors and GIS has been developed and presented in the following section. This method is thethe useuse ofof IoTIoT sensorssensors andand GISGIS hashas beenbeen developeddeveloped andand presentedpresented inin thethe followingfollowing section.section. ThisThis part of a wider research framework described in Figure4. methodmethod isis partpart ofof aa widerwider researchresearch frameworkframework describeddescribed inin Error!Error! ReferenceReference sourcesource notnot found.found...

MunicipalityMunicipality LightLight MeasuresMeasures

UrbanUrban FMFM PoliciesPolicies ConditionCondition assessmentassessment HARD KPI HARD HARD KPI HARD BudgetingBudgeting && CostingCosting AsphaltAsphalt performancesperformances

LocationLocation BasedBased OpenOpen DataData GISGIS MaintenanceMaintenance PrioritizedPrioritized O&MO&M DecisionDecision SupportSupport IntegrationIntegration PriorizationPriorization StrategiesStrategies CarCar AccidentAccident ToolsTools StatisticsStatistics

CrowdsourcingCrowdsourcing DataData SOFT KPI SOFT KPI

CitizensCitizens feedbackfeedback

FigureFigure 4. 4. TheThe workflow workflowworkflow of of the the proposed proposed approach. approach.

TheThe ideaidea concernsconcernsconcerns thethe developmentdevelopmentdevelopment ofof aaa location-basedlocation-basedlocation-based decisiondecision supportsupport systemsystem (DSS),(DSS), ableable toto supportsupport decisions decisions on on operationsoperations maintenance maintenance andand and repairrepair atat thethe urbanurban level.level.level. DataData areare collectedcollectedcollected fromfromfrom differentdifferent sources:sources: hardhard keykey perforperformancemance indicatorsindicators (KPIs)(KPIs) concerningconcerning technicaltechnical andand measurablemeasurable datadata Appl. Sci. 2020, 10, 2174 8 of 30 Appl. Sci. 2020, 10, x FOR PEER REVIEW 8 of 31 diprovidingfferent sources: information hard direct key performancely related to the indicators safety in (KPIs) the urban concerning environment, technical and and soft measurable KPIs calculated data providingfrom open informationdata and the directly crowdsourcing related to campaign the safety 71, in allowing the urban us environment, to capture information and soft KPIs impacting calculated on fromsafety open issues, data even and though the crowdsourcing in an indirect campaign manner. Since [70], allowingthe described us to capturephenomena information belong to impacting the built onenvironment safety issues, and even urban though domain, in an information indirect manner. is collected Since theand describedintegrated phenomena in a GIS through belong which to the builtdata environmentprocessing and and visualization urban domain, can be information carried out and is collected relations and unveiled. integrated These in tools a GIS allow through quantitative which data and processingqualitative analysis, and visualization data mapping can be and carried representations, out and relations are include unveiled. in a location-based These tools allow DSS which quantitative can be andrelated qualitative to policies analysis, at the municipality data mapping level, and to representations, obtain informed are decisions include inon aprioritization location-based of DSSoperations, which canmaintenance be related and to policiesrepair (OM&R) at the municipality procedures. level,This process to obtain allows informed OM&R decisions to be optimized on prioritization when the of operations,resources (time, maintenance budget and and people) repair are (OM&R) scarce procedures.and have to be This optimized process and allows prioritized. OM&R to be optimized whenIn the this resources article only (time, one budget part of and the people)wider research are scarce schema and havedepicted to be in optimized Error! Reference and prioritized. source not found.In has this been article addressed, only one since part the of implementation the wider research of the schema procedures depicted and tools in Figure employed4 has for been the addressed,development since of the the location-based implementation DSS of can the be procedures carried out and gradually. tools employed This article for focuses the development on the collection of the location-basedof data related to DSS the cantechnical be carried performances out gradually. of the urban This article component focuses of lighting, on the collection data processing of data through related toGIS the and technical the development performances of data of the analysis urban and component visualization of lighting, through data GIS, processing integrating through databases GIS from and thedifferent development open sources. of data analysis and visualization through GIS, integrating databases from different open sources. 3.1. Urban Car Accidents and Fatalities: Data Availability 3.1. Urban Car Accidents and Fatalities: Data Availability The present research focus on the urban lighting related to security and safety issues and data aboutThe the present accidents. research European focus data on the are urban widely lighting available related 70 toand security report andaccident safety data issues in anddifferent data aboutcountries the underlining accidents. Europeana trend of reduction data are widely in the last available decade [ 71(the] and data report are available accident from data 2006 in ditoff 2015).erent countriesThe accidents underlining data are a aggregated trend of reduction in three in different the last decademain sectors (the data defined are available by the typologies from 2006 of to roads: 2015). The(1) inside accidents the urban data are area, aggregated (2) motorways, in three (3) di ruralfferent area main (not sectors motorways). defined The by thedata typologies show that ofthe roads: 37% (1)of the inside accidents the urban occur area, in (2) the motorways, urban areas (3) and rural the area 55% (not in motorways). rural areas. TheThe data lower show percentage that the 37%of car of theaccidents accidents occur occur on inthe the motorway. urban areas These and data the 55% are indescribing rural areas. a situation The lower in percentage which the ofproblem car accidents is not occurstrictly on related the motorway. to the speed These of datathe cars are describing(higher in athe situation motorways) in which but thethat problem could be is notreasonably strictly relatedconnected to theto speedthe distribution of the cars of (higher the territory in the motorways) infrastructure but thatand couldservices be reasonablysuch as lighting. connected Error! to theReference distribution source of not the found. territoryb,c, infrastructure respectively, andpresent services the distribution such as lighting. of female Figure and5 b,c,male respectively, roles in the presentcase of accidents. the distribution As clearly of female shown and by the male data roles presented in the case in the of two accidents. figures, As males clearly areshown generally by the datadrivers presented involved in in the the two accidents figures, (71%), males whilst are generally females have the drivers an equal involved distribution in the in accidentsthe three roles. (71%), whilst females have an equal distribution in the three roles.

(a)

Figure 5. Cont. Appl. Sci. 2020, 10, x FOR PEER REVIEW 9 of 31 Appl. Sci. 2020, 10, 2174 9 of 30

(b)

(c)

Figure 5.5.( a()a Classification) Classification of carof accidentscar accidents in Europe in Europe depending depending on road typologyon road (btypology) female distribution(b) female ofdistribution the role in accidentsof the role (driver, in accidents passenger (driver, or pedestrian), passenger or (c )pedestrian), male distribution (c) male of thedistribution role in the of accidents the role (driver,in the accidents passenger (driver, or pedestrian). passenger or pedestrian).

InIn thethe citycity adoptedadopted asas aa casecase studystudy inin thethe presentpresent research,research, datadata areare analyzedanalyzed toto correlatecorrelate thethe lightinglighting values to to the the security security of of the the roads. roads. The The city city of Milan, of Milan, Italy, Italy, has hasthus thus been been considered considered as a ascase a casestudy study and andthe theregional regional databases databases have have been been adop adoptedted as as data data repository repository about about car car accidents andand fatalities.fatalities. StatisticsStatistics proveprove that that the the car car accidents accidents occurring occurring during during the the night night are farare morefar more dangerous dangerous than thethan ones the occurringones occurring during during the day the because day because of the visualof the andvisual behavioral and behavioral problems problems of users thatof users increase that theincrease task dithefficulties task difficulties (e.g., readiness (e.g., to readiness respond toto stimuli, respond attention, to stimuli, drowsiness, attention, use drowsiness, of substances use that of altersubstances the state that of alter vigilance, the state etc.). of Thevigilance, most recentetc.). The data most provided recent bydata Lombardy provided Region by Lombardy [72] show Region that 11,604[72] show car accidents that 11,604 occurred car accidents in Milan occurred during the in year Milan 2011, during around the 18% year of 2011, these duringaroundthe 18% night of these time. Althoughduring the they night are time. a small Although percentage they of are the a wholesmall pe numberrcentage of accidents,of the whole those numb thater occurred of accidents, during those the nightthat occurred caused 22% during of the the injuries night andcaused 42% 22% of the of deathsthe injuries (Figure and6). 42% These of datathe deaths support (Error! the concept Reference that darknesssource not can found. aggravate). These the visualdata support and behavior the concept conditions that of darkness the drivers can exposing aggravate the the night visual activities and tobehavior a high levelconditions of danger of the and drivers possible exposing critical the situations. night activities to a high level of danger and possible criticalThere situations. are many reasons that explain the statistics shown in Figure6 and none of these should be neglectedThere or are underestimated. many reasons that During explain the night, the statistics for example, shown there in Error! are more Reference people source driving not under found. the eandffects none of drugs of these or alcohol, should peoplebe neglected are tired, or thereunderest are fewerimated. police During controls, the night, etc. The for inadequateexample, there lighting are conditionmore people is, nevertheless, driving under among the effects the causes of drugs of car or accidents. alcohol, Therefore,people are consideringtired, there allare these fewer issues, police a correlationcontrols, etc. could The beinadequate found between lighting car condition accidents is, occurring nevertheless, during among the night the causes hours andof car incidence accidents. of carTherefore, accidents. considering all these issues, a correlation could be found between car accidents occurring during the night hours and incidence of car accidents. Appl. Sci. 2020, 10, x 2174 FOR PEER REVIEW 10 of 3031

18%

82%

Day Night (a)

22%

78%

Day Night

(b)

42%

58%

Day Night

(c)

Figure 6. (a) Car accidents in Milan, Italy during the year 2011, classified classified according to the time of the day of the event. ( b) Injured people in car accidents in Mila Milann during 2011, classifiedclassified according to the time of the day of the accident. ( (c)) number number of of fatalities fatalities during during the car car accidents accidents classified classified according to the time of the day ofof thethe event.event.

3.2. Open-Data Analysis Since the finalfinal aim of the research concernsconcerns the integration of the proposed method with data concerning car accidents during the night hours, the available open datasets concerning the Milan municipality (Italy) have been collected and analyzed.analyzed. There are two main open datasets about car accidents in the citycity ofof Milan:Milan: thethe firstfirst isis providedprovided byby thethe Lombardy Lombardy Region Region [ 7272] and it gives yearly data about car accidents in the city of Milan from 20002000 to 2011. In this dataset, incidents are described by thethe meansmeans ofof 101101 parameters parameters regarding regarding numbers numbers of of injuries injuries and and fatalities, fatalities, time time of theof the day day when when the Appl.Appl. Sci. 2020,, 10,, 2174x FOR PEER REVIEW 1111 ofof 3031

the incident occurred, weather condition (i.e., rainy or sunny), types and number of vehicles involved incident(i.e., cars, occurred, trucks, bikes weather and bicycles), condition etc. (i.e., rainy or sunny), types and number of vehicles involved (i.e., cars,The second trucks, open bikes dataset and bicycles), 73 is provided etc. by the municipality of Milan, it offers less detailed data aboutThe the second accidents, open namely dataset only [73] isth providede number byof theincidents, municipality injuries of and Milan, fatalities, it offers but less it detailedhas monthly data aboutdata from the accidents,2001 to 2017 namely divided only by the nine number geographical of incidents, zones injuries numbered and from fatalities, 1, the but city it center, has monthly to 9 a dataperipheral from 2001zone toin 2017the north divided of Milan by nine (for geographical a representation zones ofnumbered the zones see from Error! 1, the Reference city center, source to 9 anot peripheral found.). zoneHistorical in the data north from of Milan.this dataset Historical show datathat the from number this dataset of accidents show thatdoes the not number vary very of accidentsmuch from does the notcity varycenter very to the much peripheral from the zone, city but center the tonumber the peripheral of fatalities zone, changes but the significantly. number of fatalitiesIn Zones changes5 and 7, significantly. those with more In Zones fatalities, 5 and there 7, those are withseven more times fatalities, the deaths there registered are seven in times Zone the 1, deathsthat with registered less fatalities in Zone as 1,plotted that with in Error! less fatalities Reference as plottedsource not in Figure found.7.. TheThe numbernumber ofof accidentsaccidents are are more inin ZoneZone 88 andand ZoneZone 99 withwith aa didifferencefference fromfrom ZoneZone 11 ofof 5%5% andand fromfrom Zone Zone 2 2 of of 50%. 50%.

1600 9

1400 8

7 1200

6 1000

5 800 4

600 3

400 2

200 1

0 0 Zone 1 Zone 2 Zone 3 Zone 4 Zone 5 Zone 6 Zone 7 Zone 8 Zone 9

Car Accidents Injuries Casualties

Figure 7. Total number of car accidents, injured peoplepeople andand fatalitiesfatalities inin the nine zone of Milan area in 2017.2017. TheThe left left axis axis represents represents the numberthe number of accidents of accidents and injured and injured people whilepeople the while right axisthe representsright axis therepresents number the of fatalitiesnumber of (namely, fatalities causalities). (namely, causalities).

Noteworthy, considering considering the the historical historical data data about about car car accidents accidents in in Milan, Milan, it it is is possible possible to to register register a stronga strong decrement decrement in thein the time time span span from from 2001 2001 to 2017, to 2017, as shown as shown in Figure in Error!8. Considering Reference the source mean not of thefound. nine. Considering zones, in 2017 the there mean were of lessthe nine than zones, half of in the 2017 incidents there were that occurred less than in half 2001 of (fromthe incidents a minimum that ofoccurred 45% in in Zone 2001 6 (from to a maximum a minimum of of 60% 45% in in Zone Zone 4 and6 to a an maximum average valueof 60% of in 53% Zone of 4 reduction). and an average This meansvalue of that 53% security of reduction). and safety This campaign means that produced security some and results safety in campaign the culture produced and control some of accidentsresults in inthe the culture urban and area. control of accidents in the urban area. The lessless hazardoushazardous zone, i.e., thosethose where fewer accidents occur, do not change, however, from 2001 toto 2017.2017. TheyThey areare the the Zones Zones numbered numbered 2, 2, 4, 54, and 5 and 6 (Figure 6 (Error!8). TheReference most dangerous source not zone found. in 2001,). The zone most 9, turneddangerous out zone to be in the 2001, third zone most 9, dangerousturned out into 2017,be the exceeded third most in dang numbererous of in accidents 2017, exceeded by Zones in number 8 and 3. of accidentsThe same by trendZones discussed 8 and 3. for the number of car accidents can be observed for the number of road accidents and consequent injuries between 2001 and 2017. The average decrease in the nine zones is around 56% (minimum 48% in Zone 6 and maximum 60% in Zone 9) (Figure9) and the zones where fewer injuries are recorded are the ones numbered 2, 4, 5 and 6. This is not unexpected because the number of injuries is strictly correlated with the number of car accidents. Appl. Sci. 2020, 10, x FOR PEER REVIEW 12 of 31

3000

2500

2000

Appl.1500 Sci. 2020, 10, 2174 12 of 30 Appl. Sci. 2020, 10, x FOR PEER REVIEW 12 of 31 1000 3000 500 2500 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 20162017 Zone 1 Zone 2 Zone 3 Zone 4 Zone 5 Zone 6 Zone 7 Zone 8 Zone 9 1500 Figure 8. Car accidents in Milan according to the year and zone. The left axis represents the number 1000of accidents.

500 The same trend discussed for the number of car accidents can be observed for the number of

road0 accidents and consequent injuries between 2001 and 2017. The average decrease in the nine zones is2001 around 2002 56% 2003 (minimum 2004 2005 48% 2006 in 2007 Zone 2008 6 and 2009 maximum 2010 2011 60% 2012 in 2013 Zone 2014 9) 2015(Error! 201 62017Reference source not found.Zone) 1and theZone 2zonesZone where 3 Zonefewer 4 injuriesZone 5 areZone recorded 6 Zone 7are theZone 8ones Zonenumbered 9 2, 4, 5 and 6. This is not unexpected because the number of injuries is strictly correlated with the number of Figure 8. Car accidents in Milan according to the year andand zone. The left axis represents the number car accidents. of accidents.

4000 The same trend discussed for the number of car accidents can be observed for the number of 3500 road accidents and consequent injuries between 2001 and 2017. The average decrease in the nine zones3000 is around 56% (minimum 48% in Zone 6 and maximum 60% in Zone 9) (Error! Reference source2500 not found.) and the zones where fewer injuries are recorded are the ones numbered 2, 4, 5

and2000 6. This is not unexpected because the number of injuries is strictly correlated with the number of car accidents. 1500

40001000

3500500

30000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 20162017 2500 Zone 1 Zone 2 Zone 3 Zone 4 Zone 5 Zone 6 Zone 7 Zone 8 Zone 9 2000 Figure 9. Number of people injured in a car accident in Milan according toto thethe yearyear andand thethe zone.zone. 1500

1000The analysisanalysis of thethe historicalhistorical data about fatalities in car accidentsaccidents does not lead toto similarsimilar considerationsconsiderations500 reportedreported forfor injuriesinjuries data.data. Although therethere is a positive trend in the number of fatalities, fatalities, which goes from a total of 82 casualties in 2001 to 40 in 2017, this cannot be said about the single zone which0 goes from a total of 82 casualties in 2001 to 40 in 2017, this cannot be said about the single zone where the2001 trend 2002 in 2003 the number 2004 2005 of deaths 2006 2007 seems 2008 almost 2009 a 2010 random 2011 result 2012 (Figure 2013 2014 10). 2015 20162017 The maximumZonereduction 1 Zone 2 is recordedZone 3 inZone Zone 4 1Zone where 5 anZone 80% 6 ofZone reduction 7 Zone 8 of injuriesZone 9 is registered

(due to also the restriction to vehicles in the city center introduced in 2011–2012, the so called area Figure 9. Number of people injured in a car accident in Milan according to the year and the zone. C). the minimum reduction is 13% in Zone 5. Nevertheless, the average value of reduction is 51% according to the trend of the general increase of security on the city roads. The analysis of the historical data about fatalities in car accidents does not lead to similar If the focus is shifted from yearly data to monthly data (Figure 11), it can be underlined that there considerations reported for injuries data. Although there is a positive trend in the number of fatalities, is a similar trend in all the nine zones in August, when in Milan there is a significant decrease of people which goes from a total of 82 casualties in 2001 to 40 in 2017, this cannot be said about the single zone because of the summer holidays, as the month characterized by fewer car accidents and consequently less injuries. By contrast, May and October are the months when more car accidents and injuries occur. The average value of fatalities does not follow the same schema and, as registered for the yearly values, it is not possible to unveil a marked trend or a mathematical correlation in the data shown in Figure 11c) however Zone 9 and Zone 4 have the peak values and Zone 1 the lowest fatality value with zero in July. The monthly distribution of the accidents and injuries are absolutely regular with the following progression from the most secure to the more characterized by accident and injuries zones: Zones 2, 6, 5, 4, 7, 1, 8, 3, 9. Appl. Sci. 2020, 10, x FOR PEER REVIEW 13 of 31 where the trend in the number of deaths seems almost a random result ( 20

18

16

14

12

10

8

6

4

2

0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 20162017

Zone 1 Zone 2 Zone 3 Zone 4 Zone 5 Zone 6 Zone 7 Zone 8 Zone 9

). The maximum reduction is recorded in Zone 1 where an 80% of reduction of injuries is registered (due to also the restriction to vehicles in the city center introduced in 2011–2012, the so called area C). Appl.the Sci.minimum2020, 10, 2174reduction is 13% in Zone 5. Nevertheless, the average value of reduction is13 of51% 30 according to the trend of the general increase of security on the city roads.

20

18

16

14

12

10

8

6

4

2

0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 20162017

Zone 1 Zone 2 Zone 3 Zone 4 Zone 5 Zone 6 Zone 7 Zone 8 Zone 9 Appl. Sci. 2020, 10, x FOR PEER REVIEW 14 of 31 FigureFigure 10.10. NumberNumber ofof fatalitiesfatalities inin carcar accidentaccident inin MilanMilan accordingaccording toto yearyear andand zone.zone.

If the focus is shifted from yearly data to monthly data (Error! Reference source not found.), it can be underlined that there is a similar trend in all the nine zones in August, when in Milan there is a significant decrease of people because of the summer holidays, as the month characterized by fewer car accidents and consequently less injuries. By contrast, May and October are the months when more car accidents and injuries occur. The average value of fatalities does not follow the same schema and, as registered for the yearly values, it is not possible to unveil a marked trend or a mathematical correlation in the data shown in Error! Reference source not found.c) however Zone 9 and Zone 4 have the peak values and Zone 1 the lowest fatality value with zero in July. The mo(anthly) distribution of the accidents and injuries are absolutely regular with the following progression from the most secure to the more characterized by accident and injuries zones: Zones 2, 6, 5, 4, 7, 1, 8, 3, 9.

(b)

(c)

FigureFigure 11.11. AverageAverage valuesvalues overover 1616 yearsyears ofof thethe numbernumber ofof carcar accidentsaccidents ((aa),), numbernumber of of injuriesinjuries ( (bb)) andand numbernumber ofof fatalitiesfatalities (c(c)) in in each each month month of of the the year year for for the the nine nine zones zones in in Milan. Milan.

Historical data have been integrated into GIS maps for further analyses, and Error! Reference source not found. represents the average values of car accidents, injured people and fatalities/casualities from 2001 to 2017 in the 9 areas of Milan. As can be easily inferred, the level of aggregation of the information represented in Error! Reference source not found. is not sufficient to be employed for carrying out a reliable correlation with the collected data concerning the light intensity in some district of the city. Therefore, the integration of this data with the method presented should be considered as a next step of development of the present research. A higher level of granularity of the data is needed to perform the planned steps of the research and detailed data have already been requested by the authors to the municipality and traffic police, however, they are not available as open data. Appl. Sci. 2020, 10, 2174 14 of 30

Historical data have been integrated into GIS maps for further analyses, and Figure 12 represents the average values of car accidents, injured people and fatalities/casualities from 2001 to 2017 in the 9 areas of Milan. As can be easily inferred, the level of aggregation of the information represented in Figure 12 is not sufficient to be employed for carrying out a reliable correlation with the collected data concerning the light intensity in some district of the city. Therefore, the integration of this data with the method presented should be considered as a next step of development of the present research. A higher level of granularity of the data is needed to perform the planned steps of the research and detailed data have already been requested by the authors to the municipality and traffic police, however, they are notAppl. available Sci. 2020 as, 10 open, x FOR data. PEER REVIEW 15 of 31

(a)

(b)

(c)

FigureFigure 12. 12.Representation Representation of of average average car car accidents accidents ((aa),), injuredinjured people ( (bb)) and and casualties casualties (c ()c )in in Milan Milan zoneszones from from 2001 2001 to 2017.to 2017.

ForFor this this reason, reason, the the authors authors are are in in contact contact withwith thethe locallocal police authority, authority, which which could could provide provide accessaccess to moreto more detailed detailed data data on on car car accidents. accidents. The The mostmost criticalcritical Zone 9 9 of of Milan Milan has has been been chosen chosen and and usedused as anas an experimental experimental field field of of the the illumination illuminationmeasurements, measurements, and data data about about car car accidents accidents with with higherhigher granularity granularity in thein the same same Zone Zone 9 will 9 will be adopted,be adopted, as received,as received, as aas source a source of dataof data for for the the next next step of thestep research of the research to evaluate to evaluate the specific the specific correlation correlation between between urban urban lighting lighting and and security, security, following following an ongoingan ongoing research research line which line which is in theis in process the process of being of being deepened deepened [74 ].74.

3.3. Internet of Things (IoT) Light Sensor The data collection and analysis method should be considered as the core part of this article. Light data collection required a light meter that can be interfaced easily with a GIS-based informative system. Given the variable nature of the urban lighting, it is important to design a measurement system able to make easy the monitoring and the storage of the data and supporting a frequent update of the data. As demonstrated in literature 75, the IoT paradigm represents the ideal solution for the interconnection of sensors to remote informative systems. The design requirement of such IoT sensors are: (1) High range of light detection, because the device has to measure from an almost complete dark Appl. Sci. 2020, 10, 2174 15 of 30

3.3. Internet of Things (IoT) Light Sensor The data collection and analysis method should be considered as the core part of this article. Light data collection required a light meter that can be interfaced easily with a GIS-based informative system. Given the variable nature of the urban lighting, it is important to design a measurement system able to make easy the monitoring and the storage of the data and supporting a frequent update of the data. As demonstrated in literature [75], the IoT paradigm represents the ideal solution for the interconnection of sensors to remote informative systems. The design requirement of such IoT sensors are:

(1) High range of light detection, because the device has to measure from an almost complete dark environment (the Moon’s reflection is about 1 lx) to a high light level under the streetlamps (the values could be near 1.000 lx); (2) Geo-located measures, in order to elaborate them in the GIS to provide the data mapping at district level in the case study zone of the city of Milan; (3) A wireless communication channel; (4) Good portability, compactness and stability allowing the sensor to be carried around the city without significant issues (humidity problems, low weight for transportability during the inspection phases for monitoring, ability to maintain a pre-established geometric set-up, position accuracy, etc.).

Accordingly, given the aforementioned requirements, a prototype has been developed and realized using an Arduino Uno board. Besides the board itself, the main parts of the light meter are listed below:

(1) A high accuracy ambient light digital 16-bit resolution sensor (Adafruit VEML7700 Lux Sensor) characterized by a 16-bit dynamic range for ambient light detection from 0 lx to about 120 klx to detect the difference of illuminance at the street level going from dark areas and detecting the punctual light collected under the perpendicular direction of the lamp; the resolution of the sensor is 2 lx, compatible with the application. (2) A 56-channel GPS module with an accuracy of 2.5 m on the horizontal position able to localize the monitoring device in the street and possibly connect the information at a detailed location level; (3) A micro SD Shield for an Arduino board equipped with a 64 GB micro SD card, required to store the measurement data collected from the ambient light sensor; (4) A wireless communication channel for data transfer to a GIS-based informative system.

Figure 13 shows the different parts of the light meter assembled for some tests before using it on the experimental field. In order to fix the GPS and light sensors in a horizontal position and to make it more portable, a cardboard box has been used to introduce the monitoring device. It is noteworthy that the cardboard is white, therefore, the light reflected from the casing may affect the illuminance detection. In a similar way, the shadows cast by the operator could affect the measured luminance values. Further implementation of the experimental set-up will be tested in the following steps of the research.

3.4. Data Collection and Geographic Information System Integration The stream of light data monitored from the IoT sensor are then pushed into the GIS-based informative system. The GIS stores the ID of the IoT sensors, the light measurement data, the time stamp of the measurement, and the GPS position. Potentially, more than one IoT sensor can push the data into the informative system, enabling the use of a network of IoT sensor for data mapping of the urban area. The interaction between the IoT sensor, the GIS system and the elaboration is shown in Figure 14.

Appl. Sci. 2020, 10, 2174 16 of 30

Figure 13. The light meter assembled from an Arduino Uno board without the mechanical envelope. The main components of the board are highlighted in the picture.

The IoT system has been equipped with a GPS sensor, allowing it to capture the location of each light intensity measure. Location data are collected in the WGS 84 (EPSG 3857) coordinate system. The accuracy of the GPS sensor (2.5 m) is enough for collecting the location of the light intensity measurement on a map. However, this accuracy could be affected by issues caused by the so-called urban canyons and issues generated by surveying areas characterized by trees with high foliage (e.g., boulevards, parks, etc.). which may determine a lower reliability in the registration of the geographic coordinates [76]. Therefore, for this research, the monitored area has been divided in a grid of 10 m × 10 m squares (20 m2 area each) and an algorithm computing the arithmetical average of the detected light intensity for each square has been developed. The details of the algorithm are described in Section 5.4. The results of this computation allow us to identify the brightest and darkest areas in the urban environment, avoiding issues related to accuracy of the GPS signal. The 20 m2 grid allows us to overcome the issues related to the precision in the detection of the position of the urban elements, while correctly representing the streets and the paths in pedestrian areas.

Biomolecules 2020, 10, 399; doi:10.3390/biom10030399 www.mdpi.com/journal/biomolecules Appl. Sci. 2020, 10, x FOR PEER REVIEW 17 of 30

boulevards, parks, etc.). which may determine a lower reliability in the registration of the geographic coordinates 76. Therefore, for this research, the monitored area has been divided in a grid of 10 m × 10 m squares (20 m2 area each) and an algorithm computing the arithmetical average of the detected light intensity for each square has been developed. The details of the algorithm are described in Section 5.4. The results of this computation allow us to identify the brightest and darkest areas in the urban environment, avoiding issues related to accuracy of the GPS signal. The 20 m2 grid allows us Appl.to overcome Sci. 2020, 10 the, 2174 issues related to the precision in the detection of the position of the urban elements,17 of 30 while correctly representing the streets and the paths in pedestrian areas.

Figure 14. Interface between the Geographic Information System (GIS)-based informative system and the Internet of Things (IoT) light sensor.

4. CaseFigure Study 14. Interface between the Geographic Information System (GIS)-based informative system and the Internet of Things (IoT) light sensor. The proposed methods have been applied to a district in Zone 9 used as case study in the city of Milan,4. Case Italy. Study The “Città Study” neighborhood which is located around the technical university pole of the city was chosen as test bed for the implementation of the survey campaign. The proposed methods have been applied to a district in Zone 9 used as case study in the city of An area of approximately 2000 m2 was surveyed in two nights of measurements (15 July and Milan, Italy. The “Città Study” neighborhood which is located around the technical university pole 23 July 2019). The area was chosen, among other reasons, because of the high average occurrence of of the city was chosen as test bed for the implementation of the survey campaign. car accidents and injured people from 2001 to 2017 as shown in Section 3.1. The surroundings of the An area of approximately 2000 m2 was surveyed in two nights of measurements (15 July and 23 Politecnico di Milano campus showed a concentration of degradation of the urban components and, July 2019). The area was chosen, among other reasons, because of the high average occurrence of car sometimes, of social distress (Figure 15). accidents and injured people from 2001 to 2017 as shown in Section 3.1. The surroundings of the Moreover, the Politecnico di Milano university campus is currently undergoing a series of Politecnico di Milano campus showed a concentration of degradation of the urban components and, refurbishments and regeneration interventions, involving the retrofit of some existing buildings, sometimes, of social distress (Error! Reference source not found.). the construction of new ones and the renovation of public open spaces that will lead to a new configuration of the area. Thus, an analysis of the current light quality in the area will be useful data for the designers involved in the renovation project. As described in the previous section, the IoT sensor has been designed in order to optimize its compactness and portability. Each of the IoT sensors can be mounted on a mobile mechanical structure for tracking the level of luminosity on the road. During the case study, the sensor has been mounted on an odometer to allow easy monitoring. An operator performed the recognition on the neighborhood area involved in the case study using the odometer equipped with the light sensor. A single IoT sensor was used to collect all data during the two nights of the duration of the experiment. Each measurement campaign generated 2000 sampled data. It should be highlighted that the IoT sensor can be mounted on any mobile mechanical structure, such as an unmanned aerial vehicle (UAV) or unmanned ground vehicle (UGV), to create an autonomous measuring system for the mapping of an entire city. Such an approach has the great advantage that the IoT sensors are monitoring a predefined and well-known path. Notwithstanding such benefits, the integration of the light sensors on autonomous monitoring system is out of the scope of the current research work and it will be investigated in future research work. Appl.Appl. Sci. Sci. 20202020,, 1010,, x 2174 FOR PEER REVIEW 1818 of of 30 30

(a)

(b)

(c)

Figure 15. Cont. Appl. Sci. 2020, 10, x 2174 FOR PEER REVIEW 19 of 30

(d)

(e)

(f)

Figure 15. Cont. Appl. Sci. 2020, 10, x 2174 FOR PEER REVIEW 20 of 30

(g)

Figure 15. PicturesPictures of of critical critical areas areas (a) (toa– (g)g) where buildin buildingg and and street street degradation, degradation, lack lack of lighting of lighting and andperceived perceived unsecurity unsecurity in the inneighborhood the neighborhood were detect wereed detected and in which and inthe which proposed the approach proposed was approach tested was tested. Moreover, the Politecnico di Milano university campus is currently undergoing a series of 5.refurbishments Results and regeneration interventions, involving the retrofit of some existing buildings, the construction of new ones and the renovation of public open spaces that will lead to a new 5.1. Degradation Assessment configuration of the area. Thus, an analysis of the current light quality in the area will be useful data for theThe designers measurement involved campaign in the renovation allowed to project. collect information about the situation of the district as firstAs asdescribed description in the of previous the degradation section, (e.g.,the IoT physical sensor degradation has been designed of buildings in order and to roads,optimize social its distresscompactness of people and mobileportability. temporarily Each of living the IoT solutions sensors such can as be camper mounted vans, on use a mobile of the city mechanical parks as hygienicstructure supportfor tracking services the level for temporary of luminosity living, on th discontinuitye road. During of illuminationthe case study, of the sensor district has ranging been frommounted illumination on an odometer to complete to allow darkness) easy monitoring spots in the. area.An operator performed the recognition on the neighborhoodThe district area has involved some hotspots in the suchcase study as pubs us anding the restaurants odometer with equipped a low concentration, with the light whilstsensor. the A mainsingle vocation IoT sensor of was the zoneused isto residential. collect all data Community during the services two nights of the ofdistrict the duration are a sportsof the groundexperiment. and someEach measurement small urban parks. campaign The measurementsgenerated 2000 have sampled been data. validated It should by double be highlighted checking thethat collected the IoT valuessensor incan the be two mounted monitoring on any dates. mobile mechanical structure, such as an unmanned aerial vehicle (UAV)The or social unmanned situation ground has been vehicle varied (UGV), in the two to datescreate and an theautonomous mobile houses measuring changed system their positions for the nearmapping to the of park an entire to a non-monitored city. Such an location.approach Duringhas the the great measurement advantage campaignthat the IoT it was sensors possible are tomonitoring verify the a neighborhoodpredefined and conditions well-known where path. buildings Notwithstanding degradation such due benefits, to lack the of maintenanceintegration of and the cleaninglight sensors procedures on autonomous (Figure 15monitoringa,b), road system problems is out of of the the pavement scope of andthe cu ofrrent the horizontalresearch work signage and forit will pedestrians be investigated (Figure in 15futureb,d,f), research asymmetrical work. lighting and temporary mobile housing in camper vans (Figure 15c–e) and excessively darkness for people safety and street security (Figure 15g) have been5. Results detected.

5.2.5.1. Degradation Urban Lighting Assessment Assessement The Europeanmeasurement standard campaign EN 13201-3 allowed [77 to] collect gives rules information on the points about wherethe situation lighting of measurements the district as shouldfirst as bedescription taken. Since of thethe purposedegradation of survey (e.g.,was physical not the degradation measurement of ofbuildings the photometric and roads, quality social of thedistress lighting of people installations, mobile duringtemporarily the data living collection solutions the such grids as mandatedcamper vans, by theuse Europeanof the city standardparks as werehygienic not used.support Points services located for intemporary the shade living, of objects discontinuity (i.e., trees) thatof illumination should be avoided of the accordingdistrict ranging to the standardsfrom illumination were, in to fact, complete preferred darkness) measurement spots in locations the area. because the step passage from dark shadow to shinyThe lightingdistrict has can some affect hotspots driving performance.such as pubs and restaurants with a low concentration, whilst the mainMore vocation than of 2000 the pointszone is distributed residential. along Community almost 6.5 services km were of the collected district in are the a first sports campaign. ground Halfand ofsome them, small reported urban inparks. Figure The 16 a,measurements were below 32 have lx and been more validated than 52% by ofdouble the points checking were the below collected 50 lx, whichvalues arein the the two light monitoring levels of public dates. areas with dark surroundings according to the NOAO (National Appl. Sci. 2020, 10, x FOR PEER REVIEW 21 of 30

The social situation has been varied in the two dates and the mobile houses changed their positions near to the park to a non-monitored location. During the measurement campaign it was possible to verify the neighborhood conditions where buildings degradation due to lack of maintenance and cleaning procedures (Figure 15a,b), road problems of the pavement and of the horizontal signage for pedestrians (Figure 15b,d,f), asymmetrical lighting and temporary mobile housing in camper vans (Figure 15c–e) and excessively darkness for people safety and street security (Figure 15g) have been detected.

5.2. Urban Lighting Assessement The European standard EN 13201-3 77 gives rules on the points where lighting measurements should be taken. Since the purpose of survey was not the measurement of the photometric quality of the lighting installations, during the data collection the grids mandated by the European standard were not used. Points located in the shade of objects (i.e., trees) that should be avoided according to the standards were, in fact, preferred measurement locations because the step passage from dark shadow to shiny lighting can affect driving performance. More than 2000 points distributed along almost 6.5 km were collected in the first campaign. Half Appl. Sci. 2020, 10, 2174 21 of 30 of them, reported in Error! Reference source not found.a, were below 32 lx and more than 52% of the points were below 50 lx, which are the light levels of public areas with dark surroundings according Opticalto the NOAO Astronomy (National Observatory), Optical Astronomy national center Observat for ground-basedory), national night-time center for astronomy ground-based in the Unitednight- Statestime astronomy [78]. Moreover, in the it United is worth States noting 78. thatMoreover, 12.5% of it theis worth points noting are above that500 12.5% lx and of the three points of them are exceedabove 500 2500 lx and lx. Thesethree of points them where exceed recorded 2500 lx. Thes undere points the direct where light recorded of lighting under systems the direct of privatelight of parkinglighting lotsystems thus theyof private are not parking significant lot thus in the they number are not of significant records and in the they number are not of describing records and specific they areasare not with describing high illuminance specific areas level with but high points illuminance of discontinuity level but in points the average of discontinuity distribution. in the Thus, average such valuesdistribution. can be Thus, disregarded such values in the can following be disregarded analysis. in the following analysis.

(a)

(b)

Figure 16. (a) Histogram of the distribution of lux level of the measurement points; (b) classification of the detected illuminance according to NOAO thresholds.

A further parameter that could be added in the lighting evaluation could be the uniformity ratio given by the average illuminance and the minimum illuminance (Eav/Emin) that has defined values ranging from 3 (for expressway to 6 for local roads) and it is a crucial value for visual comfort and adaptability in the scotopic vision of the human eye. The problems that could be related to the luminance values are:

(1) The illuminance must be sufficiently high in relation to the dimensions of the object, to the contrasts, to the time of perception and to the age of the subject; (2) The luminance of the area immediately adjacent to the visual task must not be higher than the visual task itself; (3) The contrast between the immediate background of the visual task and the environment must not be accented and the transition should be gradual; (4) The most appropriate solutions have been taken to avoid glare. Appl. Sci. 2020, 10, x FOR PEER REVIEW 23 of 31

Figure 16. (a) Histogram of the distribution of lux level of the measurement points; (b) classification of the detected illuminance according to NOAO thresholds.

A further parameter that could be added in the lighting evaluation could be the uniformity ratio given by the average illuminance and the minimum illuminance (Eav/Emin) that has defined values ranging from 3 (for expressway to 6 for local roads) and it is a crucial value for visual comfort and adaptability in the scotopic vision of the human eye. The problems that could be related to the luminance values are: (1) The illuminance must be sufficiently high in relation to the dimensions of the object, to the contrasts, to the time of perception and to the age of the subject; (2) The luminance of the area immediately adjacent to the visual task must not be higher than the visual task itself; (3) The contrast between the immediate background of the visual task and the environment must Appl. Sci. 2020, 10, 2174 22 of 30 not be accented and the transition should be gradual; (4) The most appropriate solutions have been taken to avoid glare. It isis usefuluseful to to remember remember that that illuminance illuminance describes describes the measurementthe measurement of the of amount the amount of light of falling light ontofalling and onto illuminating and illuminating a given a surface given area.surface Illuminance area. Illuminance is also correlated is also correlated with how with humans how perceivehumans theperceive brightness the brightness of an illuminated of an illuminated area. area. As aa result,result, mostmost people people use use the the terms terms illuminance illuminance and and brightness brightness interchangeably; interchangeably; however, however, this canthis leadcan lead to misunderstanding, to misunderstanding, as brightness as brightness is also is also used used to define to de luminance.fine luminance. Illuminance Illuminance refers refers to a specificto a specific kind ofkind light of measurement, light measurement, while brightness while brightness refers to the refers visual to perceptions the visual andperceptions physiological and sensationsphysiological of light.sensations of light. Luminance describes the measurement of the amountamount of light emitting, passing through or reflectedreflected fromfrom aa particular particular surface surface from from a solida solid angle. angle. It alsoIt also designates designates how how much much luminous luminous power power can becan perceived be perceived by the by human the human eye. This eye. means This thatmeans luminance that luminance indicates indicates the brightness the brightness of light emitted of light or reflectedemitted or from reflected a surface from (Figure a surface 17). (Error! Reference source not found.).

Figure 17. Definition and differences between (a) illuminance in lux (lm/m2) and (b) luminance in Figure 17. Definition and differences between (a) illuminance in lux (lm/m2) and (b) luminance in cd/m2 (lm/sr m2). × cd/m2 (lm/sr × m2). 5.3. Urban Lighting and Safety Perception 5.3. Urban Lighting and Safety Perception Current LED road lighting has a poor longitudinal luminance uniformity which is described as a “zebraCurrent effect” LED [79] thatroad can lighting easily has cause a poor eye fatiguelongitudinal and the luminance reduction uniformity of safe driving which (Figure is described 18). as a “zebraTo solveeffect” this 79 that problem, can easily some cause researchers eye fatigue adopted and the specific reduction illuminance of safe driving distribution (Error! Reference curves to accomplishsource not found. a certain). illuminance and luminance uniformity by adjusting the curve parameters [80]. But theseTo solve methods this cannotproblem, achieve some ideal researchers illuminance adopted and luminance specific illumi uniformitynance simultaneouslydistribution curves and areto onlyaccomplish appropriate a certain for illuminance specific roads. and The luminance main aspects uniform affityecting by adjusting the visual the perception curve parameters in road driving80. But andthese thus methods that arecannot relevant achieve for ideal security illuminance and safety and luminance for drivers unif areormity average simultaneously road surface and luminance, are only overallappropriate and longitudinalfor specific roads. luminance The main uniformity, aspects glareaffecting control, the visual and color perception rendering in road index driving [81]. However, and thus thethat pedestrianare relevant perception for security of securityand safety is alsofor drivers a combination are average of environmental road surface andluminance, social degradation overall and elementslongitudinal and luminance signs as well uniformity, as the aforementioned glare control, and lighting color rendering conditions index of the 81. road However, and sidewalks. the pedestrian perception of security is also a combination of environmental and social degradation elements and signs 5.4.as well Urban as the Illuminance aforementioned Mapping lighting conditions of the road and sidewalks. Points have been imported and georeferenced in a GIS (QGIS) and have been used for the development of the data analysis algorithm. For this purpose, a polygon grid 10m 10m has × been developed considering an extension equal to the maximum extent of the coordinated points representing the measured light intensity. The first insight of where the measurement characterized by higher light intensity are localized in the selected area is shown in Figure 19. Data have been categorized according to 6 classes of light intensity (shown in the legend of Figure 19), considering equal intervals. The last class shows values characterized by a very high illuminance values despite it is composed by only 2 values, recorded under direct lights (a private parking lot and the very bright sign of a commercial activity). Appl. Sci. 2020, 10, 2174 23 of 30 Appl. Sci. 2020, 10, x FOR PEER REVIEW 23 of 30

(a)

(b)

(c)

(d)

Figure 18. (a) “Zebra effect” on the pavement (horizontal bands with different lighting levels); (b) safety perception of a pedestrian related to lighting levels and degradation signs on the built environment (lack of maintenance and vandalism on the facades); (c) difference of pavements in the city and increased danger for bikes and motorbikes when raining; (d) shading of the urban lighting provided by the trees distributed in parking areas and urban residual parks decreasing visual capability and safety perception for pedestrians. Appl. Sci. 2020, 10, 2174 24 of 30

It is possible to note an unevenness of distribution of the illuminance values: the most illuminated Appl. Sci. 2020, 10, x FOR PEER REVIEW 24 of 30 locations are those where interesting activities are spotted (the university main fronts, some leisure locationsFigure such 18. ( asa) “Zebra pub and effect” restaurants on the pavement and the crossing(horizonta ofl bands some with bigger different urban lighting streets). levels); By contrast, (b) a largesafety part perception of the measurements of a pedestrian fall related in the to low lighting level oflevels illumination and degradation range, betweensigns on the 0 and built 539 lx. Thus,environment unequally spaced(lack of intervalsmaintenance are and considered vandalism (shown on the infacades); the legend (c) difference of Figure of 20pavements) in order into the better analyzecity theand visualincreased effect danger of brightness for bikes and and motorbikes darkness when that can raining; be perceived (d) shading by of citizens. the urban The lighting raw light 2 measurementprovided databy the are trees spatially distributed merged in in parking a grid ofar 10eas m and10 urban m squares residual (an parks area of decreasing 20 m each), visual in order × to reducecapability the number and safety of perception points to analyze for pedestrians. on the GIS. In addition, as discussed in the previous section, the measurement set-up is affected by a variability contribution which depend by the mechanical system5.4. Urban used Illuminance to support Mapping light sensors during the measurement campaign through the city. Although suchPoints variability have is been well belowimported the accuracyand georeferenced required by in the a analysisGIS (QGIS) performed and have in thisbeen paper, used the for light the datadevelopment in each square of the ofdata the analysis grid are algorithm. arithmetically For this averaged, purpose, to a reduce polygon the grid measurement 10m × 10m variability. has been Andeveloped SQL (structured considering query an language) extension script equal allows to th toe computemaximum the arithmeticalextent of the average coordinated light intensity, points maximum and minimum light intensity values and the number of points in each of 10 m 10 m representing the measured light intensity. The first insight of where the measurement characterized× polygonalby higher light grid. intensity Figure 20 are represents localized the in outcomesthe selected of thisarea data is shown analysis. in Error! The visualization Reference source allows not us tofound. identify. Data areas have in been the test categorized bed neighborhood according with to 6 diclassesfferent of illuminance light intensity levels. (shown The in lower the legend ranges of luminosityError! Reference correspond source to locationsnot found. where), considering it could be equal useful intervals. to intensify The the last illumination class shows level values and tocharacterized provide new by systems a very high to increase illuminance safety values and security despite in it a is developing composed areaby only of the 2 values, city subjected recorded to regenerationunder direct lights processes. (a private parking lot and the very bright sign of a commercial activity).

Figure 19. Results of the measurement campaign. The dimens dimensionion of the circles is proportional to light intensity. SixSix lightlight intervalsintervals (equally(equally spaced)spaced) areare consideredconsidered inin thethe figure.figure.

It is possible to note an unevenness of distribution of the illuminance values: the most illuminated locations are those where interesting activities are spotted (the university main fronts, some leisure locations such as pub and restaurants and the crossing of some bigger urban streets). By contrast, a large part of the measurements fall in the low level of illumination range, between 0 and 539 lx. Thus, unequally spaced intervals are considered (shown in the legend of Error! Reference source not found.) in order to better analyze the visual effect of brightness and darkness that can be perceived by citizens. The raw light measurement data are spatially merged in a grid of 10 m × 10 m squares (an area of 20 m2 each), in order to reduce the number of points to analyze on the GIS. In addition, as discussed in the previous section, the measurement set-up is affected by a variability Appl. Sci. 2020, 10, x FOR PEER REVIEW 25 of 30

contribution which depend by the mechanical system used to support light sensors during the measurement campaign through the city. Although such variability is well below the accuracy required by the analysis performed in this paper, the light data in each square of the grid are arithmetically averaged, to reduce the measurement variability. An SQL (structured query language) script allows to compute the arithmetical average light intensity, maximum and minimum light intensity values and the number of points in each of 10m × 10m polygonal grid. Error! Reference source not found. represents the outcomes of this data analysis. The visualization allows us to identify areas in the test bed neighborhood with different illuminance levels. The lower ranges of luminosity correspond to locations where it could be useful to intensify the illumination level and to Appl.provide Sci. 2020 new, 10 systems, 2174 to increase safety and security in a developing area of the city subjected25 of to 30 regeneration processes.

Figure 20.20. MeasuredMeasured illuminanceilluminance (average (average lux), lux), aggregated aggregated within within a grid a grid of squares of squares of 10 of m 10 10m m× each.10 m × each. 6. Discussion 6. DiscussionSeveral metrics can be used to infer quantitatively the perceived level of security of citizens in the diSeveralfferent metrics area of acan city: be crimes used to per infer area, quantitatively road accidents, the personalperceived assault, level of and security many of more. citizens There in isthe evidence different from area theof a literature city: crimes that per such area, metrics road accidents, are directly personal affected assault, by the and quality many of streetmore. lights:There pooris evidence lighting from is correlated the literature with that an increasesuch metrics of the are perceived directly affected level of insecurity.by the quality However, of street it lights: is not possiblepoor lighting to increase is correlated indiscriminately with an increase the level of of the lighting perceived in any level area of of insecurity. the city, both However, for economical it is not aspossible well as to environmental increase indiscriminately reasons (energy the level consumption of lighting and in any light area pollution). of the city, Thus, both it isfor important economical to haveas well an as accurate environmental map of thereasons quality (energy of lighting consumption in the diandfferent light areas pollution). of the Thus, city. Such it is important data should to behave correlated an accurate with map other of metricsthe quality to identify of lighting the areain the which different requires areas an of improvementthe city. Such indata quality should light. be Currentcorrelated research with other work metrics focuses to on identify a methodology the area ablewhich to requires generate an an improvement accurate mapping in quality of lighting light. qualityCurrent in research urban area. work These focuses data on are a methodology correlated with able road to generate accidents an information accurate mapping to demonstrate of lighting the feasibilityquality in ofurban the proposedarea. These approach. data are correlated with road accidents information to demonstrate the feasibilityThe survey of the campaignproposed approach. for quality light monitoring has been carried out in the most accurate way possible,The survey despite campaign some drawbacks for quality that light can monitoring be highlighted. has been Due carried to the out experimental in the most natureaccurate of way the proposedpossible, despite procedure, some the drawbacks orthogonality that of can the be sensor highlighted. to the ground Due to could the experimental be improved, ensuringnature of thatthe the detection tools are always horizontal. Moreover, the variation of height from the ground could have been affected for the same reasons, despite always carefully remained within a reasonable range. These issues may cause a bias in the detection of the light intensity data. However, the final evaluations are carried out on average values computed on the 10 m 10 m grid, therefore the single values are × not considered, but in intermediate steps of the proposed approach. A further improvement of the research concerns the accuracy in the detection of the position of the sensor when recording the light intensity data. For the case study presented in this article a grid 10 m 10 m has been adopted. This allows us to represent the light intensity along the streets and × the open spaces of the neighborhood, although not providing enough detail needed to represent and analyses specific technical units of the urban environment (e.g., carriageway, pavement, parking lots Appl. Sci. 2020, 10, 2174 26 of 30 etc.). This aspect will be addressed and some research insights will be carried out in order to be able to reach a finer granularity in the analysis and visualization of the geolocated data. An oversampling of the data from ambient length sensors could be a possible solution. The proposed method has been developed following the guidelines of the EN 13201-4 [82], in particular, the following steps of the research on the orthogonality of the light detecting sensor will be improved as well as the continuity in the height at which the measures are taken. However, the standard should be used when the purpose of characterizing lighting installations considers the photometric quality parameters, an objective that is not completely within the scope of this research, oriented towards the assessment of the security of urban space related to the light intensity. Another issue that may have affected the detection of the light data concerns the setup of the Arduino equipped with the sensors. The breadboard has been inserted in a white paperboard box. The light reflecting on the paperboard may have affected the detection of the light intensity, decreasing the accuracy of the values. This issue could be overcome adopting an improved installation setup, or simply employing a black paperboard box.

7. Conclusions The planning of the renovations of street light infrastructures in urban areas should take into consideration several parameters. The level of light in urban areas cannot be arbitrarily high due to economical as well as environmental (energy consumption and light pollution) reasons. At the same time, the level, and the quality, of light is strictly correlated with the perceived level of security of the citizen. The research work proposed a methodology for the planning of light renovations based on the correlation of GIS data from different sources. The map of the quality of the street light in the different urban areas is correlated with other GIS data, which can be considered quantitative indexes of the perceived level of security. In particular, in the current research the number and geographical distribution of road accidents was considered an example of metrics to be used to validate the proposed approach. The publicly available road accident data of the Milan area has been analyzed. It is possible to underline as a further step of the research that it will be enriched with the road accidents data that will allow a significant comparison with drivers’ difficulties in the tested area of the city located in Zone 9 which is the most subject to car accidents. The idea is to extend the indicators (the hard KPI) with the crowdsourcing data about perceived safety from the citizens (soft KPI) which can compose a structured proposal for implementing the lighting systems of the neighborhoods and provide the municipality an effective strategic plan to develop the renovated university campus area. It is worth underlining that the use of only the measures suggested by the EN 13201-4 are not enough to assess urban security because lighting not directly related to roads are needed for this purpose to include the pedestrians’ perception under the different segments of the city areas (including green areas and urban parks). The urban lighting data mapping could drive informed decisions on increasing the public lighting infrastructure since public light can improve social security and perception of physical degradation, increasing community pride and participation to the social good and urban care. The increase of lighting points could be developed with new smart systems able to use renewable energies and dimmer light related to the presence sensors avoiding a waste of energy and recreating a customized living lighting environment which could envision an augmented user experience and environmental and economic sustainability levels. The proposed GIS integration is a powerful tool for urban regeneration, driving urban policies at the local scale with an efficient granularity of intervention that could increase the quality of the urban environment.

Author Contributions: Conceptualization, L.C.T. and F.R.C.; methodology L.C.T., F.R.C, S.R.; software, F.R.C, N.M., P.B.; validation, S.R., F.R.C. and P.B.; formal analysis, N.M.; investigation, F.R.C., N.M., L.C.T.; resources, F.R.C.; data curation, N.M., F.R.C; writing—original draft preparation, L.C.T, N.M.; writing—review and editing, S.R., L.C.T.; visualization, F.R.C., N.M.; supervision, L.C.T., F.R.C, A.L.C.C.; project administration, L.C.T; funding acquisition, F.R.C. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Appl. Sci. 2020, 10, 2174 27 of 30

Acknowledgments: The authors would like to thank Marco Pasetti for the useful discussion about the monitoring setup in the planning phase. Conflicts of Interest: The authors declare no conflict of interest.

References

1. UNFPA-United Nations Population Fund. State of World Population 2018: The Power of Choice; UNFPA: New York, NY, USA, 2018; p. 156. ISBN 978-1-61800-032-3. 2. UNFPA-United Nations Population Fund. State of World Population 2007: Unleashing the Potential of Urban Growth; UNFPA: New York, NY, USA, 2007; p. 108. ISBN 978-0-89714-807-8. 3. IPCC. Working Group III (Mitigation)–Chapter 12. Human Settlements, Infrastructure and Spatial Planning. Available online: https://www.ipcc.ch/site/assets/uploads/2018/02/ipcc_wg3_ar5_chapter12.pdf (accessed on 19 March 2020). 4. UNFPA-United Nation Population Fund. Available online: http://www.unfpa.org/pds/urbanization.htm (accessed on 19 March 2020). 5. Hamidi, S.; Ewing, R.; Tatalovich, Z.; Grace, J.B.; Berrigan, D. Associations between Urban Sprawl and Life Expectancy in the United States. Int. J. Environ. Res. Public Health 2018, 15, 861. [CrossRef] 6. Rodwin, V.G.; Gusmano, M.K. Growing Older in World Cities: New York, London, Paris, and Tokyo; Vanderbilt University Press: Nashville, Tennessee, 2006; ISBN 0-8265-1490-1. 7. STAP The Scientific and Technical Advisory Panel of the Global Environment Facility. Sustainable Urbanization Policy Brief: Proliferation of Urban Centres, their Impact on the World’s Environment and the Potential Role of the GEF; Report to the 5th GEF Assembly, México May 2014’; Global Environment Facility: Washington, DC, USA, 2014. 8. C40 Cities Climate Leadership Group. Available online: http://c40.org (accessed on 20 March 2020). 9. The Global Commission on the Economy and Climate. The New Climate Economy Report. 2014, Better Growth Better Climate, Charting a New Path or Low Carbon Growth and Safer Climate; New Climate Economy (NCE): Washington, US, USA, 2014. 10. Rizzi, M.; Depari, A.; Ferrari, P.; Flammini, A.; Rinaldi, S.; Sisinni, E. Synchronization Uncertainty Versus Power Efficiency in LoRaWAN Networks. IEEE Trans. Instrum. Meas. 2019, 68, 1101–1111. [CrossRef] 11. Rinaldi, S.; Bonafini, F.; Ferrari, P.; Flammini, A.; Sisinni, E.; Cara, D.D.; Panzavecchia, N.; Tine, G.; Cataliotti, A.; Cosentino, V.; et al. Characterization of IP-Based communication for smart grid using software-defined networking. IEEE Trans. Instrum. Meas. 2018, 67, 2410–2419. [CrossRef] 12. Pasetti, M.; Rinaldi, S.; Manerba, D. A virtual power plant architecture for the demand-side management of smart prosumers. Appl. Sci. 2018, 8, 432. [CrossRef] 13. Rinaldi, S.; Della Giustina, D.; Ferrari, P.; Flammini, A. Distributed monitoring system for voltage dip classification over distribution grid. Sustain. Energy Grids Netw. 2016, 6, 70–80. [CrossRef] 14. Shlomo, A. Making Room for a Planet of Cities, Policy Focus Report; Lincoln Institute of Land Policy: Cambridge, MA, USA, 2011; ISBN 978-1-55844-212-2. 15. Huchzermeyer, M.; Karam, A. Informal Settlements: A Perpetual Challenge; University of Cape Town Press: Cape Town, South Africa, 2007; ISBN 1-9197-1394-8. 16. Un-Habitat. United Nations Human Settlements Programme. In Enhancing Urban Safety and Security: Global Report on Human Settlements 2007; Earthscan: London, UK, 2007; ISBN 971-1-84407-475-4. 17. IPCC. Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change; Edenhofer, O.R., Pichs-Madruga, Y., Sokona, E., Farahani, S., Kadner, K., Seyboth, A., Adler, I., Baum, S., Brunner, P., Eickemeier, B., et al., Eds.; Cambridge University Press: Cambridge, UK, 2014. 18. Hale, J.D.; Daies, G.; Fairbrass, A.J.; Matthews, T.J.; Rogers, C.D.F.; Sadler, J.P. Mapping Lightscape: Spatial Patterning of Artificial Lightinh in a Urban Landscape. PLoS ONE 2013, 8, e61460. [CrossRef][PubMed] 19. Wanvik, P.O. Effects of road lighting: An analysis based on Dutch accident statistics 1987–2006. Accid. Anal. Prev. 2009, 41, 123–128. [CrossRef] 20. Uddin, M.; Huynh, N. Truck-involved crashes injury severity analysis for different lighting conditions on rural and urban roadways. Accid. Anal. Prev. 2017, 108, 44–55. [CrossRef] Appl. Sci. 2020, 10, 2174 28 of 30

21. Zhang, Y.; Lu, H.; Qu, W. Geographical detection of traffic accidents spatial stratified heterogeneity and influence factors. Int. J. Environ. Res. Public Health 2020, 17, 572. [CrossRef] 22. Gould, M.; Poulter, D.R.; Helman, S.; Wann, J.P. Judgments of approach speed for motorcycles across different lighting levels and the effect of an improved tri-headlight configuration. Accid. Anal. Prev. 2012, 48, 341–345. [CrossRef] 23. Wood, J.M.; Isoardi, G.; Black, A.; Cowling, I. Night-time driving visibility associated with LED streetlight dimming. Accid. Anal. Prev. 2018, 121, 295–300. [CrossRef][PubMed] 24. Bozorg, S.; Tetri, E.; Kosonen, I.; Luttinen, T. The Effect of Dimmed Road Lighting and Car Headlights on Visibility in Varying Road Surface Conditions. LEUKOS-J. Illum. Eng. Soc. North Am. 2018, 14, 259–273. [CrossRef] 25. Setyaningsih, E.; Putranto, L.S.; Soelami, F.N. Analysis of the visual safety perception and the clarity of traffic signs and road markings in the presence of road lighting in straight and curved road. MATEC Web Conf. 2018, 181, 04001. [CrossRef] 26. Fotios, S. A review of design recommendations for P-class road lighting in European and CIE documents—Part 1: Parameters for choosing a lighting class. Lighting Res. Technol. 2019.[CrossRef] 27. Chawuthai, R. Monitoring roadway lights and pavement defects for nighttime street safety assessment by sensor data analysis and visualization. Sens. Mater. 2018, 30, 2267–2279. [CrossRef] 28. Longcore, T.; Rich, C. Ecological light pollution. Front. Ecol. Environ. 2004, 2, 191–198. [CrossRef] 29. Bruce-White, C.; Shardlow, M. A Review of the Impact of Artificial Light on Invertebrates; Buglife: Peterborough, UK, 2011; p. 32. 30. Gallaway, T.; Olsen, R.N.; Mitchell, D.M. The economics of global light pollution. Ecol. Econ. 2010, 69, 658–665. [CrossRef] 31. The European PPP Expertise Centre (EPEC). Energy Efficient Street Lighting: Guidelines; World Bank Group: Washington, DC, USA, 2010. 32. Mohamed, S. Smart street lighting control and monitoring system for electrical power saving by using VANET. Int. J. Commun. Netw. Syst. Sci. 2013, 6, 351–360. [CrossRef] 33. Ozadowicz, A.; Grela, J. Energy Saving in the street lighting control system—A new approach based on the EN-15232 standard. Energy Effic. 2017, 10, 563. [CrossRef] 34. Crisp, V.H.C.; Henderson, G. The energy management of artificial lighting use. Lighting Res. Technol. 1982, 14, 193–206. [CrossRef] 35. Elvidge, C.; Baugh, K.; Hobson, V.; Kihn, E.; Kroehl, H.; Davis, E.; Cocero, D. Satellite inventory of human settlements using nocturnal radiation emissions: A contribution for the global toolchest. Glob. Chang. Biol. 1997, 3, 387–395. [CrossRef] 36. Elvidge, C.D.; Cinzano, P.; Pettit, D.R.; Arvesen, J.; Sutton, P.; Small, C.; Nemani, R.; Longcore, T.; Rich, C.; Safran, J.; et al. The Nightsat mission concept. Int. J. Remote Sens. 2007, 28, 2645–2670. [CrossRef] 37. Sutton, P.C.; Taylor, M.J.; Anderson, S.; Elvidge, C.D. Sociodemographic Characterization of Urban Areas Using Nighttime Imagery, Google Earth, Landsat and “Social” Ground Truthing; Weng, Q., Quattrochi, D., Eds.; Urban Remote Sensing; CRC Press: Boca Raton, FA, USA, 2007; pp. 291–310. 38. Levin, N.; Duke, Y. High spatial resolution night-time light images for demographic and socio-economic studies. Remote Sens. Environ. 2012, 119, 1–10. [CrossRef] 39. Barducci, A.; Marcoionni, P.; Pippi, I.; Poggesi, M. Effects of light pollution revealed during a nocturnal aerial survey by two hyperspectral imagers. Appl. Opt. 2003, 42, 4349–4361. [CrossRef][PubMed] 40. Kuechly, H.U.; Kyba, C.C.; Ruhtz, T.; Lindemann, C.; Wolter, C.; Fischer, J.; Hölker, F. Aerial survey and spatial analysis of sources of light pollution in Berlin, Germany. Remote Sens. Environ. 2012, 126, 39–50. [CrossRef] 41. Wu, J.G.; Jenerette, G.D.; Buyantuyev, A.; Redman, C.L. Quantifying spatiotemporal patterns of urbanization: The case of the two fastest growing metropolitan regions in the United States. Ecol. Complex. 2011, 8, 1–8. [CrossRef] 42. Forman, R.T.T.; Godron, M. Landscape Ecology; Wiley: New York, NY, USA, 1986; p. 619. 43. Schneider, A.; Woodcock, C.E. Compact, dispersed, fragmented, extensive? A comparison of urban growth in twenty-five global cities using remotely sensed data, pattern metrics and census information. Urban Stud. 2008, 45, 659–692. Appl. Sci. 2020, 10, 2174 29 of 30

44. Doll, C. Thematic Guide to Night-time Light Remote Sensing and Its Applications; CIESIN: New York, NY, USA, 2008; p. 4. 45. Luginbuhl, C.B.; Lockwood, G.W.; Davis, D.R.; Pick, K.; Selders, J. From The Ground Up I: Light Pollution Sources in Flagstaff, Arizona. Publ. Astron. Soc. Pac. 2009, 121, 185–203. [CrossRef] 46. ILE. Guidance Notes for the Reduction of Obstructive Light; The Institution of Lighting Engineers: Warwickshire, UK, 2005. 47. Morgan-Taylor, M. Light Pollution and Nuisance: The Enforcement Guidance for Light as a Statutory Nuisance. J. Plan. Environ. Law 2006, 8, 1114–1127. 48. Flanders Government. Order of the Flemish Government of 1 June 1995 concerning General and Sectoral provisions relating to Environmental Safety: Chapter 4.6; Control of nuisance due to light; Flanders Government: Brussels, Belgium, 1995. 49. Fernandes Carvalho, D.; Ferrari, P.; Sisinni, E.; Depari, A.; Rinaldi, S.; Pasetti, M.; Silva, D. A test methodology for evaluating architectural delays of LoRaWAN implementations. Pervasive Mob. Comput. 2019, 56, 1–17. [CrossRef] 50. CIE International Commission on Illumination. CIE 194:2011: On Site Measurement of the Photometric Properties of Road and Tunnel Lighting; Peter Blattner: Hochdorf, Switzerland, 2011; ISBN 978 3 901906 92 3. 51. Leccese, F.; Salvadori, G.; Rocca, M. Critical analysis of the energy performance indicators for road lighting systems in historical towns of central Italy. Energy 2017, 138, 616–628. [CrossRef] 52. Stevens, R.G. Light-at-night, circadian disruption and breast cancer: Assessment of existing evidence. Int. J. Epidemiol. 2009, 38, 963–970. [CrossRef][PubMed] 53. Cinzano, P.; Falchi, F.; Elvidge, C.D. The first World Atlas of the artificial night sky brightness. Mon. Not. Royal Astron. Soc. 2001, 328, 689–707. [CrossRef] 54. Mizon, B. Light Pollution Responses and Remedies; Springer: London, UK, 2002; p. 216. 55. Imhoff, M.L.; Lawrence, W.T.; Stutzer, D.C.; Elvidge, C.D. A technique for using composite DMSP/OLS “city lights” satellite data to map urban area. Remote Sens. Environ. 1997, 61, 361–370. [CrossRef] 56. Small, C.; Pozzi, F.; Elvidge, C.D. Spatial analysis of global urban extent from DMSP-OLS night lights. Remote Sens. Environ. 2005, 96, 277–291. [CrossRef] 57. Land SAT. Available online: https://www.sciencedirect.com/topics/earth-and-planetary-sciences/landsat (accessed on 19 March 2020). 58. Sutton, P.; Roberts, D.; Elvidge, C.; Baugh, K. Census from Heaven: An estimate of the global human population using night-time satellite imagery. Int. J. Remote. Sens. 2001, 22, 3061–3076. [CrossRef] 59. Hölker, F.; Moss, T.; Griefahn, B.; Kloas, W.; Voigt, C.C.; Henckel, D.; Hänel, A.; Kappeler, P.M.; Völker, S.; Schwope, A.; et al. The Dark Side of Light: A Transdisciplinary Research Agenda for Light Pollution Policy. Ecol. Soc. 2010, 15, 13. [CrossRef] 60. Massey, P.; Foltz, C.B. The Spectrum of the Night Sky over Mount Hopkins and Kitt Peak: Changes after a Decade. Publ. Astron. Soc. Pac. 2000, 112, 566–573. [CrossRef] 61. Davoudian, N. Urban Lighting for People: Evidence-Based Lighting Design for the Built Environment; RIBA Publishing: London, UK, 2018; ISBN 978-1859468210. 62. Lyytimaki, J.; Tapio, P.; Assmuth, T. Unawareness in environmental protection: The case of light pollution from traffic. Land Use Policy 2012, 29, 598–604. [CrossRef] 63. Falchi, F.; Cinzano, P.; Elvidge, C.D.; Keith, D.M.; Haim, A. Limiting the impact of light pollution on human health, environment and stellar visibility. J. Environ. Manag. 2011, 92, 2714–2722. [CrossRef] 64. Cinzano, P. Recent progresses on a second world atlas of the night-sky brightness, LPTRAN/LPDART realistic models, tomography of light pollution, accurate validation methods and extended satellite data analysis. In Proceedings of the Starlight 2007 Conference, La Palma, Spain, 19–20 April 2007. 65. Night Earth map. Available online: www.nightearth.com (accessed on 24 June 2019). 66. Streetlight-EPC Project, Triggering the market uptake of Energy Performane Contracting through street lighting refurbishment, Co-funded by the Intelligent Energy Europe Programme of European Union. Available online: http://www.streetlight-epc.eu/ (accessed on 19 March 2020). 67. Beccali, M.; Bonomolo, M.; Leccese, F.; Lista, D.; Salvadori, G. On the impact of safety requirements, energy prices and investment costs in street lighting refurbishment design. Energy 2018, 165, 739–759. [CrossRef] 68. Lv, Z.; Li, X.; Zhang, B.; Wang, W.; Zhu, Y.; Hu, J.; Feng, S. Managing Big City Information Based on WebVRGIS. IEEE Access 2016, 4, 407–415. [CrossRef] Appl. Sci. 2020, 10, 2174 30 of 30

69. Li, M.; Lenzen, M.; Keck, F.; McBain, B.; Rey-Lescure, O.; Li, B.; Jiang, C. GIS-Based Probabilistic Modeling of BEV Charging Load for Australia. IEEE Trans. Smart Grid 2019, 10, 3525–3534. [CrossRef] 70. Re Cecconi, F.; Moretti, N.; Dejaco, M.C.; Maltese, S.; Tagliabue, L.C. Community involvement in urban maintenance prioritization. In Proceedings of the 2017 AEIT International Annual Conference, Cagliari, Italy, 20–22 September 2017; pp. 1–6. [CrossRef] 71. European Commission. Annual Accident Report 2017; European Road Safety Observatory: Bruxelles, Belgium, 2017; Available online: www.erso.eu (accessed on 20 March 2020). 72. Incidenti stadali Comune di Milano, Opena Data. Available online: https://www.dati.lombardia.it/Sicurezza/ INCIDENTI-STRADALI-nel-COMUNE-MILANO/y7qm-hfqy (accessed on 19 March 2020). 73. Portale Open Data Comune di Milano. Available online: http://dati.comune.milano.it/dataset/ds177- trafficotrasporti-incidenti-stradali-persone-infortunate-mese-zona (accessed on 19 March 2020). 74. Kumar, S.; Deshpande, A.; Ho, S.S.; Ku, J.S.; Sarma, S.E. Urban Street Lighting Infrastructure Monitoring Using a Mobile Sensor Platform. IEEE Sens. J. 2016, 16, 12. [CrossRef] 75. Bellagente, P.; Ferrari, P.; Flammini, A.; Rinaldi, S. Adopting IoT framework for Energy Management of Smart Building: A real test-case. In Proceedings of the 2015 IEEE 1st International Forum on Research and Technologies for Society and Industry, Turin, Italy, 16–18 September 2015; RTSI 2015-Proceedings. pp. 138–143. 76. Cui, Y.; Ge, S.S. Autonomous vehicle positioning with gps in urban canyon environments. IEEE Trans Robot Autom. 2003, 19, 15–25. 77. EN 13201-3:2015: Road lighting-Part 3: Calculation of performance; NSAI Standard: Dublin, Ireland, 2015. 78. NOAO – National Optical Astronomy Observatory, Recommended Light Levels (Illuminance) for Outdoor and Indoor Venues. Available online: https://www.noao.edu/education/QLTkit/ACTIVITY_Documents/ Safety/LightLevels_outdoor+indoor.pdf (accessed on 19 March 2020). 79. Hu, X.; Qian, K. Optimal design of optical system for LED road lighting with high illuminance and luminance uniformity. Appl. Opt. 2013, 52, 5888–5893. [CrossRef] 80. Feng, Z.; Luo, Y.; Han, Y. Design of LED freeform optical system for road lighting with high luminance/illuminance ratio. Opt. Express 2010, 18, 22020–22031. [CrossRef] 81. Schreuder, D.A. Road Lighting for Safety; Inst of Civil Engineers Pub: London, UK, 1998. 82. European Committee for Standardization (CEN) CEN-EN 13201–4: 2015-Methods of Measuring Lighting Performance; European Committee for Standardization: Brussels, Belgium, 2015.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).