applied sciences Article Towards a Hybrid Approach to Context Reasoning for Underwater Robots Xin Li *, José-Fernán Martínez and Gregorio Rubio Research Center on Software Technologies and Multimedia Systems for Sustainability (CITSEM), Campus Sur, Technical University of Madrid, 28031 Madrid, Spain; [email protected] (J.-F.M.); [email protected] (G.R.) * Correspondence: [email protected]; Tel.: +34-914-524-900 (ext. 20794) Academic Editor: Antonio Fernández-Caballero Received: 23 December 2016; Accepted: 8 February 2017; Published: 15 February 2017 Abstract: Ontologies have been widely used to facilitate semantic interoperability and serve as a common information model in many applications or domains. The Smart and Networking Underwater Robots in Cooperation Meshes (SWARMs) project, aiming to facilitate coordination and cooperation between heterogeneous underwater vehicles, also adopts ontologies to formalize information that is necessarily exchanged between vehicles. However, how to derive more useful contexts based on ontologies still remains a challenge. In particular, the extreme nature of the underwater environment introduces uncertainties in context data, thus imposing more difficulties in context reasoning. None of the existing context reasoning methods could individually deal with all intricacies in the underwater robot field. To this end, this paper presents the first proposal applying a hybrid context reasoning mechanism that includes ontological, rule-based, and Multi-Entity Bayesian Network (MEBN) reasoning methods to reason about contexts and their uncertainties in the underwater robot field. The theoretical foundation of applying this reasoning mechanism in underwater robots is given by a case study on the oil spill monitoring. The simulated reasoning results are useful for further decision-making by operators or robots and they show that the consolidation of different reasoning methods is a promising approach for context reasoning in underwater robots. Keywords: context reasoning; uncertainty; Multi-Entity Bayesian Network (MEBN); underwater robots; ontology; context awareness 1. Introduction Context awareness [1] is important in many research domains, such as Smart Cities [2], Smart Homes [3], Ambient Assisted Living [4], and Smart Grids [5]. As a key enabler for entities to understand their environment and make adaptations accordingly, context awareness implies an effective exploitation of contexts. With advances in sensing, pervasive computing, and communication technologies, more and more contexts can be obtained and promisingly used. To make the most of the available contexts is key to achieving context awareness. In general, three conventional approaches have been used to achieve context awareness [6]: (1) each application or domain acquires, processes, and employs contexts of its interest in its own manner; (2) some libraries that provide methods to process contexts are used in context-aware applications or domains; and (3) a context-aware framework/middleware/intermediation architecture is adopted to provide common functionalities to manage contexts and deliver context awareness. According to Li et al. [7], the third approach is regarded as the best solution due to its ability to decrease the complexity of building context-aware applications. The European Smart and Networking Underwater Robots in Cooperation Meshes (SWARMs) [8] project includes introducing context awareness into the field of underwater robotics as one of its objectives. This project aims to expand the use of unmanned underwater vehicles (e.g., AUVs or Appl. Sci. 2017, 7, 183; doi:10.3390/app7020183 www.mdpi.com/journal/applsci Appl. Sci. 2017, 7, 183 2 of 20 The European Smart and Networking Underwater Robots in Cooperation Meshes (SWARMs) [8] project includes introducing context awareness into the field of underwater robotics as one of its Appl.objectives. Sci. 2017 This, 7, 183 project aims to expand the use of unmanned underwater vehicles (e.g., AUVs2 of or 20 ROVs) in maritime and offshore operations in a collaborative, cooperative, and context-aware ROVs)manner. in The maritime SWARMs and offshoreapproach operations is underpinned in a collaborative, by designing cooperative, a semantic and and context-aware distributed manner.middleware. The SWARMsThe middleware approach islayer underpinned has been by des designingigned to a semanticfacilitate and communication distributed middleware. between Theheterogeneous middleware vehicles layer has and been a Command designed to & facilitate Control communicationStation (C & CS) between and provide heterogeneous a set of common vehicles andservices. a Command A context-aware & Control framework, Station (C & dedicated CS) and provide to delivering a set of commoncontext awareness services. A in context-aware underwater framework,robots, is designed dedicated as part to of delivering the middleware. context The awareness context-aware in underwater framework robots, aims isto designedprovide different as part ofcontext the middleware. treatments that The comply context-aware with the frameworkgeneral lifecy aimscle of to context provide awareness. different contextAs depicted treatments in Figure that 1, complythe context with theawareness general lifecyclelifecycle ofis context summarized awareness. into As four depicted essential in Figure phases1, the [9], context namely awareness context lifecycleacquisition is summarized (obtaining necessary into four context essential data), phases context [9], namely modeling context (representing acquisition contexts (obtaining in a necessarymachine- contextreadable data), and contextprocessable modeling form), (representing context reasoning contexts (deriving in a machine-readable high-level contexts and processable from available form), contextcontexts), reasoning and context (deriving dissemination high-level (distributing contexts from useful available contexts). contexts), Ontologies, and context which are dissemination considered (distributingthe most promising useful modeling contexts). technique Ontologies, in terms which of are expressiveness considered theand most interoperability promising [10], modeling have techniquebeen adopted in terms in the of context-aware expressiveness framework and interoperability to formally [10 represent], have been heteroge adoptedneous in the contexts context-aware that are frameworkexchanged tobetween formally vehicles, represent such heterogeneous as the battery contexts level of that vehicles, are exchanged capabilities between of vehicles, vehicles, turbidity, such as theposition, battery speed, level salinity, of vehicles, and capabilitieswind direction of vehicles, [11]. It is turbidity, worth noting position, that speed,operators salinity, or vehicles and wind are directionmore willing [11]. to It isuse worth high-level noting thatcontext operators informatio or vehiclesn instead are of more raw willing context to to use conceive high-level and context offer informationcontext-aware instead services of raw[12]. contextFor instan to conceivece, a piece and of offerhigh-level context-aware context, servicesvehicle A [12might]. For collide instance, with avehicle piece B of soon high-level, is more context, meaningfulvehicle for A might operators collide to with make vehicle decisions B soon, isthan more the meaningful basic context for information operators to makevehicle decisions A is out of than trajectory the basic and context heading information in the directionvehicle of vehicle A is out B of. Therefore, trajectoryand context heading reasoning in the direction based ofon vehicle information B. Therefore, formatted context in ontologies reasoning basedis import on informationant for realizing formatted context in ontologiesawareness is in important underwater for realizingrobots. In context particular, awareness the harsh in underwaternature of the robots. underwater In particular, environment the harsh and naturethe limitations of the underwater of sensors environmentattached to vehicles and the impose limitations more of challenges sensors attached in making to vehicles effective impose context more reasoning challenges in underwater in making effectiverobots. In context the underwater reasoning environment, in underwater the robots. majority In of the contexts, underwater namely environment, sensory data, the are majority prone ofto contexts,be uncertain namely [12]. sensory Uncertainties, data, are as prone an inherent to be uncertain characteristic [12]. Uncertainties, of context, as should an inherent be handled characteristic in the ofreasoning context, phase. should Therefore, be handled there in the is reasoning a need fo phase.r context-aware Therefore, thereframework is a need to forprovide context-aware suitable frameworkaccommodation to provide for the suitable inclusion accommodation of uncertainties. for the inclusion of uncertainties. Figure 1. The lifecycle of context awareness. A lotlot ofof context context reasoning reasoning methods methods have have been been off-the-shelf off-the-shelf and reviewedand reviewed in [13, 14in]; [13,14]; however, however, none of themnone of is them versatile is versatile and could and individually could individually solve all solve the reasoningall the reasoning requirements requirements in underwater in underwater robots. Thererobots. is There a necessity is a necessity for a shift for a towards shift towards a combination a combination of different of different context context reasoning reasoning methods methods [7]. Ontological[7]. Ontological