Comparative Advantages of System on Chips in Intelligent Traffic System

Home , System on a chip

Research Article: Open Access source arecredited. Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and Copyright: Received: France, E-mail:[email protected] Technology, Srinagar,India;Departmentof Automatic ControlandRobotics,EcoleCentraledeNantes, *Corresponding author: Mohammad Mujtahid,Departmentof Information Technology,NationalInstitute of Introduction 2 1 Mohammad Mujtahid Traffic System Advantages ofSystemonChipsinIntelligent Comparative forts inthisfield[1 GOOGLE andUBERprogramsarethemostvisible ef- of programslikeDARPAGrandChallenges.Currently accomplishment ofthisdreamstartedwiththeadvent nary streetsontheirown.Pioneeringstepstowards the began dreaming ofcarssmartenoughtonavigateordi- 1960s. Backthenenthusiastsofartificialintelligence(AI) systems inthefieldoftransportationstartedback and trafficconditions.Effortsfordevelopingautomated and communicationnetworkstoimprovevehicles,roads much emphasisonITS.Itutilizesthecomputerresource of theprimaryworriesbigcities,ICTprogramslay Department ofAutomaticControlandRobotics,EcoleCentraledeNantes,France Department ofInformationTechnology,NationalInstituteIndia J RobotEng3:004 Citation: With overcrowded roads and traffic jams being one With overcrowdedroadsandtrafficjamsbeingone VIBGYOR December28,2017: Mujtahid M, Sofi SA, Bashir T (2018) Comparative Advantages of System on Chips in Intelligent Traffic System. Int 21 Mjai M e a. hs s n pnacs atce itiue udr h trs f h Cetv Commons Creative the of terms the under distributed article open-access an is This al. et M, Mujtahid 2018 © System onchip,Intelligenttrafficsystem,Automation Keywords augmented includes, which ITS, of vehicle systemandintelligentdrivingplatform. development for B Model 2 Pi Raspberry of implementation the on analysis this base We microcontrollers. and (SBCs) Computers Board Single ITS to for compared suitable more of are SoCs insight why analyze an we paper provides this In study ICT. of its aspects required So all almost expanse. technological vast a over spreads program ICT parts of dynamic most the of one being ITS (ITS). System Traffic Intelligent of automation in SOCs of impact the analyzing at aim we paper this In etc. Coprocessors or modules Wi-Fi GPUs, power, processing higher peripherals, advanced for providing by microcontrollers of limitations the beyond break SOCs (PSoCs). Chips on System Programmable or (SoCs) chips on system of by overtaken assistance the for providing in wonders human kind. With more advances inthe field of automationthe microcontrolleraspect is nowbeing doing already is automation and (IoT) computing, Things (ICT) of embedded Internet Technology With niches. Communication safer and and better smarter, Information building at aimed of programs, backbone the is technology Automation Abstract ]. 1,2* Accepted:February15,2018: , ShabirASofi 1 andTariq Bashir Published: cessing power and reduced size, energy usage etc. Some cessing powerandreducedsize,energyusageetc.Some made tocreateacomputationaldevicewithhighpro - native. limiting parametersSOCsprovidedforanoptimalalter - using microcontrollersbecamedifficult.Totraverse these and machinelearningtheimplementationofautomation for processingpowerincreased.Alsowithmoredatasets the increaseinnumberofsensinginputsrequirement gracefully beenhandledbymicrocontrollers.Butwith processing andreactingforalongtimehasbeenquiet propriate commandsforreacting.Thisjobofsensing, needed toacceptsenseddata,processitandprovideap- fining automationcomponentsofITS,processorsare Robtic Enger Inter From technological perspective various efforts were From technologicalperspectivevariouseffortswere With sensing,processingandreactingbeingthede- February17,2018 natiol Jur 1 Mujtahid etal.IntJRobotEng2018,3:004 nal of ISSN: 2631-5106 Mujtahid et al. Int J Robot Eng 2018, 3:004 ISSN: 2631-5106 | • Page 2 of 9 •

Figure 1: Technology paradigm [4]. of the notable efforts in this regard are: Chips on board, [5]. Also complicated systems like audio players and au- system in package and also system on chip. With chips tomation based devices are being built upon SoCs. Also on board being simple integration of chips on a printed added on PWM modules to SoCs make it simple for pro- circuit board (PCB), system in package (SiP) is integra- grammers to manipulate signals by varying duty cycles tion of various ICs in a single package and SoC is the in- of a PWM without writing any software component for tegration of the whole system on a single chip. SiP found this task [5]. its application in performing almost all electronic system Summing up the point, SoCs have many advantages functions especially mobile phones and music players, over microcontrollers and general purpose computers. whilst system on chip (SoC) revolutionized all compu- Other than above mentioned properties some of the no- tational aspects requiring high processing in lower space, table advantages are: for example, mobile phones, Internet of Things (IoT), ICT etc. [2,3]. • Size: Many functions and features in much reduced package. In Figure 1 we have a tabular comparison of various technological paradigms and the basic scope of improve- • Flexibility: It is hard to find any parallel to SoCs in ment for them. terms of the flexibility they provide for board size, power and efficiency. Functionalities and Advances Provided by SOCS • Price Factor: SoCs provide various inbuilt functional- Flexibility in interfacing, better power efficiency and ities and ease in many aspects like interfacing. Hence ability to reconfigure hardware are some of the advan- SoCs save much of effort and time, which in turn re- tages provided by SoCs. A technological innovation to duces developmental expenditure. reduce the size of devices for integrated systems is what is provided by SoC architecture. In SoCs an entire system • Configurability: In SoCs a single pin can be config- resides on a single chip. The plug and play interfaces pro- ured to perform various tasks. This feature is very use- vided by SoCs make them more convenient for system ful whilst taking into consideration PCB design. designing. Other than this, SoCs from power perspective • Ease of Hardware of access: SoCs mostly use Linux are much efficient devices, since they are void of PCB operating systems and provide easy interfaces for tracks and burdening capacitances. Also these factors hardware support. Hence making life easy for pro- lead to overall improved system speed [5]. SoCs also programmers by providing hardware access with easy vide interface for high current requirement peripherals user interface. viz, motors, servos, stepper motors etc. Interfacing a sensor to a microcontroller or a general purpose computer Architecture is quiet complex, whilst interfacing sensors to a SoC is a In this project we have developed an intelligent ve- much simplified task. hicle model which would aid in ITS. This project is SoC SoCs have enabled us to produce much reduced foot- based. We have used Raspberry Pi for accomplishing this prints for wearable and smartphone applications e.g. project. The architecture of the project has three main touch sensors such as sliders and wheels can be connect- components: Sensing component, processing component ed directly to SoCs without using any external circuitry and implementing component.

Sensing component includes ultrasonic sensor and Pi TRIG, ECHO and GND. VCC is the 5 V input power and camera. While processing component includes Raspber- GND is the ground pin. TRIG pin is to produce the ul- ry Pi 2 and implementing component includes Motors trasonic pulse output, whilst ECHO receives the reflected and L298N driven Dual H-Bridge. input. From Raspberry Pi we generate a short 10 µs pulse to trigger input to start the ranging. Sensor sends out an Ultrasonic sensor provides the distance between vehi- 8 cycle 40 KHz ultrasonic burst and raises its echo line cle and the obstacle, while Pi camera helps in color detec- to high. Once it detects an echo, it lowers the echo line. tion for traffic signal sensing. Also it helps in processing Thus echo line is a pulse proportional to distance up to image data for providing further more inputs for auto- the object. The module can be triggered at a very fast rate mation. The image processing is performed using open of 20 cycles per second. We need to provide a 50 ms sleep CV and python. time, in order to ensure that the ultrasonic beep has faded Raspberry Pi is a small, lightweight and powerful ARM away and will not cause a false echo on next ranging [8]. based SoC. Its amazing graphic capabilities and available HDMI video output provide it an upper hand for multi- Pi camera is 5 MPixel and connects to Raspberry Pi media applications. Raspberry Pi version 2 model B, the via a ribbon connector to the CSI connector. This com- one used in our project, has a 900 MHz quad-core ARM bination of Pi camera and raspberry Pi is overwhelm- cortex A53 processor and 1 GB RAM. It has extended 40 ingly fast, having the capacity to send 1080 p images at pins for interfacing and is also provided with 4 USB ports 30 frames per second or at higher frame rates for lower and an Ethernet port. Software’s relying upon multicore resolution images. The video and still image quality for Pi processing can speed up 4X, whilst for a multi-thread cam is better than USB cams of same price. It is void of IR friendly code it can go up to 7.5X. It is a magnetic piece filter and hence can see near-IR wavelengths (700-1000 of technology for innovative programmers [6]. nm), with a tradeoff of poor color rendition. This 5 MPix- el camera has fixed-focus and supports VGA 90, 720p60 The operating system (Raspbian) used, is a free work- and 1080p30 video modes and still captures as well. It can ing framework based on Debian. Working framework be through MMAL and V4L APIs and also there are third implies the fundamental utilities and programs that run party libraries like Pi Camera Python that have been built the system on which it is installed. It is designed for Rasp- for it [9]. berry Pi module. Raspbian Operating System is designed to enhance the speed of the Raspberry Pi based on the H-Bridges are typically used in controlling motor improved instruction execution speed of ARM processor speed and direction, but can also be used for other proj- CPU of Raspberry Pi [7]. The Ultrasonic sensor used is ects. An H-Bridge is a circuit that can drive a current in HC-SR04. This sensor usually is for distance measure- either polarity and can be controlled using Pulse Width ment for distances from 2 - 400 cm. It has four pins: VCC, Modulation (PWM). The H-Bridge used is driven by

Figure 2: Hardware architecture of the project.

L298N chip. Its logical voltage requirement is 5 V and the the decided location. So after planning to act to reach the drive voltage ranges from 5 V - 35 V, though it is usually target we have to perform motor control. In motor con- marked as 12 V. It has various pins for control signal in- trol we need to control speed, direction and turning phe- put from the processor and output signal to the motors nomenon. Once we reach our target we start from sens- [10]. ing again and continue the process in a recursive fashion. These components are interfaced together for getting This schema of the software architecture is represent- to the desired result. The drive voltage provided via Dual ed in Figure 3. H-Bridge to the motors is derived from an external bat- Algorithm Overview: Since we are using “sense-plan- tery attached to the circuitry. act” module for the implementation of the project, so its Figure 2 represents the hardware setup of the afore- various steps in correspondence are: mentioned application. Sensing Environment (Sense) - In this part we simply Software Architecture: The Project has been im- use the camera and ultrasonic sensors to collect the data. plemented on the very conventional software model of “sense-plan-act”. To sense the environment, we use sen- Sensor Data Fusion (Sense) - In this part we combine sors and in the project we are using camera and ultrason- the sensor inputs, so the required action can be taken ac- ic sensor as environment understanding sensors. So the cording to the priority. model has a multi-sensor sensing module. These sensor Trajectory Generation (plan) - This is an input based inputs are fused for proper perception of the environ- module of the working model. It detects the objects in ment around. the motion path and accordingly forms the motion tra- Once the sensors have sensed the surroundings and jectory. the fusion of their data in perceivable form is received we Motion Planning (plan) - Once we have the environ- now need to plan our course of action according to the ment map, we plan the motion of the vehicle in this seg- data. In planning we perform two tasks; ment of mapping. Mapping: We create a logical equivalent of the sur- Motor Control (act) - In this part we program the mo- rounding environment from the sensed data. tors to take the required action, be it stopping, turning or Setting Target (Mission): Here we according to the accelerating on the basis of the planned trajectory. logical map decide our target position. Per-segment Planning (plan) - We use this method to Once done with planning we need to act so as to improve the efficiency of the algorithm. Figure 4 explains achieve the target. In this scenario our target is to reach the advantage of using per-segment planning.

Figure 3: Software architecture.

Figure 4: Per-segment planning.

In Figure 4 we demonstrate the overlapping segment In 2017 the introduction of DRS360 autonomous breakup. Moving by planned trajectory in segment 1, with driving platform by Mentor is an approach towards lev- a segment 1 only vision would result in the vehicle coming el five automation using centralized fusion of raw data. too close to the obstacle. Hence segments are overlapped In contrast to the previous distributed computing based and the trajectory is re-planned at the entry of segment 2. approach, the centralized raw data based approach uses Evaluation the entirety of all the acquired data and hence is capa- ble of generating a more optimal view of the vehicle en- Microcontrollers have been the most dominant el- vironment. Using raw and unfiltered data is the ideal ement in the field of augmentation of driving vehicles. method to obtain a complete, spatially and temporally With a typical ford vehicle having 25-35 electronic con- synced view of the vehicular environment, and achieving trol units (ECUs), whilst a luxury car like BMW 7 se- this target was made possible by bringing forth together ries has 60-65 ECUs [11]. The type of microcontroller the FPGA and CPU technologies on the same platform. depends on the application, be it vehicle control, power Nearly any camera, LIDAR, radar or other sensor types train or driver information. Some examples of microcon- can be adapted to the platform, which is open and trans- trollers used in automobiles are Tri-core microcontrol- parent. Mentor’s sensor fusion and object recognition lers (assembled in over 50 automotive brands), AVR mi- algorithms are advanced, but companies can apply their crocontrollers (with inbuilt ADC), PIC microcontrollers own algorithms and even chose their own chips and SoCs and Renesas microcontrollers (Latest in the automotive for the processing [12]. microcontroller family) [11]. Advanced driver assistance systems (ADAS) today employ a microcontroller with Now in a more general perspective let’s compare mi- each sensor and process the data acquired from the sen- crocontrollers and SoCs. Interfacing sensors and cameras sor, this reduces the data to be sent to the central proces- and programming them with SoCs is a much simpler task sor and also the computational load of the central proces- when compared to microcontrollers and general purpose sor. This schema of augmentation of driving vehicles has computers. Microcontrollers have their set of limitations been widely used until now. The various problems with in processing capacity and programming feasibility. this approach are, local processing of the data introduces While general purpose computers do not provide cus- delay before sending it to the central processor, problems tomizable plug and play interfaces for hardware periph- in comparing data from multiple sensors to distinguish erals like sensors and cameras. Hence SoCs provide an an aspect etc. Today engineers feel that to make these optimum mix of microcontrollers and general purpose judgements the fusion of sensor data is best handled by a computers. In Table 1, we have tried to give a basic com- central controller, which also processes the raw data from parison of microcontrollers and SoCs based on a speci- the sensors simultaneously [12]. men for each category.

Though performance and interactive operating sys- price with facilities for customizing pins to interface sen- tem provided by single board computers is better than sors and actuators. This setup is best provided by SoCs in SoCs, but they don’t provide any plug and play interfaces comparison to microcontrollers and single board com- and also they are comparatively very costly. Nor is there puters. any availability for customizing pins for interfacing pe- To interface an ultrasonic sensor with our comput- ripherals like sensors, motors etc. Hence we do not use ers and laptops is a very complicated task to accomplish. single board computers for function specific automation While interfacing ultrasonic sensors with Arduino is systems. The feasibility and low cost with a tradeoff of feasible but not simple enough. Microcontrollers do not performance by SoCs is a very optimal solution for the have an operating system, so to get our sensor integrat- function specific automation systems. Figure 5 gives a ed and running with Arduino we need to connect it to statistical analysis of instructions per unit time for vari- another computer via a serial connection and download ous raspberry Pi boards. The instruction rate of all rasp- an Arduino IDE in it. Then either upload a predefined berry Pi boards is compatible for ITS, with later versions sketch in IDE for ultrasonic sensor or write a new code being more efficient. for ultrasonic sensor to Arduino. So Arduino is not an Automation systems are based upon sensing modules interactive enough option for programming and interfac- for sensing environment using sensors, computer vision ing sensors. In contrast SoCs have feasible plug and play provided by cameras and processing platforms, central pins like microcontrollers and an interactive Linux oper- processing space that retrieves data from sensors and ating system to work upon. Raspbian operating system processes it for ambient response. The processed data contains python setup, hence providing an easy option performs response via actuators like motors, servos etc. for programming the ultrasonic sensor. With built-in li- An automation system has its specific functional domain braries for Raspberry Pi in python it is easy to configure hence it needs a moderate processing space at feasible GPIO pins for sensor interfacing. Also with the vast functionalities provided by python libraries it makes system Table 1: Comparison between SOC and microcontroller. designing easy for the programmer. Chip Microcontroller SoC The data input from the ultrasonic sensor helps us de- (Arduino) (Raspberry Pi) termine the course of action that we should take to avoid Cost Per Unit $10 $50 possible collision in case of presence of an obstacle. So we Performance 16 MHz 1.2 GHz import the data from sensors to a separate class to manip- Flexibility High High ulate the action performed by the motors. Unlike micro- Sensor Interfacing Feasible Feasible LAN Connectivity Not available Built-in Port controllers we need not create a separate sketch and then Operating System Not available Linux based and upload or write it back, rather we need to create a separate Windows 10 IOT Core program and simply import the data retrieved by sensors

Raspberry Pi with gfortran

450, 000

400, 000

350, 000

300, 000

250, 000

200, 000

150, 000

100, 000

50, 000 Thousands of Instructions per Second (KIPS)

0 Pi B (Original) Pi 2 B Pi Zero Pi 3 B Figure 5: Whetstone double precision benchmark [13].

Citation: Mujtahid M, Sofi SA, Bashir T (2018) Comparative Advantages of System on Chips in Intelligent Traffic System. Int J Robot Eng 3:004 Mujtahid et al. Int J Robot Eng 2018, 3:004 ISSN: 2631-5106 | • Page 7 of 9 • into this program. Then the output signal is carried on to thus providing us an effective length of 4 meters instead motors via Dual H-Bridge. We can manipulate the work- of 2 meters. Figure 6 explains the idea of effective length. ing of H-Bridge with the help of pulse width modulation Now maximum speed that a vehicle can attain at a (PWM). In the defined scenario the sensor data needs junction is around 120 KMPH, which is around 33 MPS. to manipulate the driving speed and direction. Driving speed is based on motor speed, and driving direction is Mathematically analyzing this situation, we have; based on relative motor speed. We can manipulate motor Vehicle speed = 33 m/s speed by varying the duty cycle of each pulse width modulation. Higher duty cycles correspond to higher speed. Time taken to travel 1 meter = 1/33 s The relative motor speed can be manipulated by setting Hence, Time taken to cover 4 meters = 4/33 s up different duty cycles for different motors, using the So the camera needs to capture the image in less than specific GPIO pins for each motor. Though the manipu- every 4/33 of a second. lation via PWM can also be provided by microcontrollers like Arduino but every-time we create a new manipula- Thus, required camera framerate > 33/4 fps = 8.25 fps. tion we have to create a new sketch and write it back on But this is not the absolute framerate of the camera, the microcontroller. However, for single board comput- it is the framerate that the camera has to maintain in si- ers we are not provided with customizable pins which can multaneity with image processing algorithms for edge be manipulated using PWM and hence we can’t perform detection, filtering etc. Also to be on the safer side we the required action. would prefer an available framerate of 10 fps. The usual Now for computer vision we use cameras and soft- framerate of a USB camera is around 15 to 20 fps with- ware platforms. In our project we have used Pi camera out any processing going on at the backend of it, and and openCV software. openCV (Open computer vision) that too for a resolution less than 1080. This framerate software is designed for aiding computer vision. Pi cam- goes plummeting down to mere 5-7 fps after applying era was also developed to aid Raspberry Pi based com- processing algorithms and that too at lower resolutions. puter vision. SoCs like Raspberry Pi made it possible to Now USB cameras with higher framerates come at very provide computer vision in low priced robots and auto- expensive prices. On the contrary Pi camera has an ab- mation devices, which otherwise was restricted to very solute framerate of 30 fps for 1080-pixel resolution and expensive automation devices alone. Now for computer a higher framerate for lower resolutions. Applying pro- vision our camera should be able to capture all the activ- cessing algorithms takes down the framerate of the cam- ity on the road, at the junction. Minimum vehicle length era to around 15 fps at minimum. Also the price of Pi is around 2 meters. So the effective length for the vehicle camera board is considerably less. Now let us analyze the to be captured by camera is 4 meters. As the camera can three available prospects of microcontroller, SoC and capture either the head of the car or the tail of the car, SBC for a practical nodal situation.

Figure 6: Vehicle effective length.

Figure 7: Future outlook of AV system [15].

Space Constraint: At a traffic junction the available (NRE) of an SoC are much lower, which save us a lot of space for implementing such devices is not so ambient maintenance cost. Parallel operations with lower clock for SBCs, whilst microcontroller and SoC setup can fit in. rates and ability to sleep when inactive reduces the power consumption [14]. These advantages have led to the use Cost Constraint: Cost of a simple and portable SBC of SoCs in many automation and smart-tech projects to is at minimum around Rs 21000 ($330). Other than that envisage a better and safer future. the cost of high-framerate USB camera is also around Rs 6400 ($100). Summing up the main amount without oth- Future Scope: The field of SoCs has yet a vast expanse er accessories like sensors PCBs etc. will be around $450. to cover. It has just started to impact the world of tech- This amount is very high compared to the implementa- nology. Smartphones, research purpose boards, hand tion price of similar functionalities using microcontrol- held devices and IoT architectures are a cogent proof of lers and SoCs. their technological excellence. Feasibility Constraint: We need a setup providing fea- SoCs can go multi-core to make their processing per- sibility for hardware connectivity and computer vision. formance even more efficient. We can let SoCs go for Computer Vision setup needs device compatibility with hyper-threading so as to make parallel processing look camera and sufficient processing power for image pro- even more convenient on the boards. This will facilitate cessing. Microcontrollers though provide easy hardware the centralized fused sensor based approach for level 5 access to sensors and have good plug and play interfaces, automation of vehicular system. but they lack compatibility for cameras and the required Moreover, the processing setup of the SoCs can be processing power hence making them not so suitable for made smart. This would enable auto turnoffs for power the situation. Whilst SBC provides very massive pro- consumption efficiency and smart switching of processor cessing power and ready camera compatibility, it lacks the easy access towards hardware and does not provide to enable temperature management and much more [4]. the luxury of plug and play interfaces. SoCs on the other We need to develop customized SoCs such that drive hand provide plug and play interfaces for easy hardware domain is separated from the interface domain. Then we access and are also very good for computer vision with need to connect the SoC to the Back-end system at the camera compatibility and enough processing power. So interface domain using a smart antenna. The connec- SoCs are an efficient and optimal choice for this case. tion to the Back-end system provides the vehicle with Now taking a look at more general advantages that information like road condition, free parking spaces etc. the SoCs can provide in our case. Better performance Whilst the segregation of drive domain and interface do- due to its capability of exploiting hardware parallelism, main is required to secure the vehicular safety even in provides us faster interfaces for retrieving and process- case of intrusion or overburdening data input from the ing sensor data. Also the non-recurring expenditures smart antenna [15].

Figure 7 gives an overview of such a customized SoC of Information Technology, National Institute of Tech- based AV system. nology Srinagar. The author also extends his gratitude towards Muhammad Talha Bilal for his assistance in im- Another important aspect to work on is the potential proving the graphic content of figures. for mixed signals. With upcoming semiconductor struc- tures we might be working on different signal levels like References 1.8 V for 28 nm process technology. We should be able 1. Jonathan Diclemente, Serban Mogos, Ruby Wang (2014) to handle this system of mixed signals. We can work on Autonomous car final report. Carnegie Melon University. generation of charge pumps for SoCs to handle such is- 2. Abhishek Ghosh (2012) System on a chip (SOC): A brief sues. This will make the AV system free to choose any details. sensor as per the requirement without worrying for sig- 3. Global Industry Analysts, Inc. (2017) “System-in-Package nal specifications [4]. (SiP) technology”. Global Industry Analysts. Hence SoCs with their current vast scope and applica- 4. Heng Sin Wei, Adrian Kong Yeng Hong, Chris Liu Chaofeng, tion also have a lively and exponentially expanding scope Goh Chee Peng, Jason Koh Sheng Fa (2012) “System on for future improvements. With companies like Renesas, chip (for mobile devices)”. NXP and Nvidia building SoCs like R Car H3, Bluebox 5. Nishant Mittal (2016) Design advantages of programmable, and Drive PX respectively and having joint efforts with efficient SoC architectures. automotive companies like, Tesla, BMW, Volvo etc. In 6. (2017) Raspberry pi technology, working and its applications. future the SoCs will play a pivotal role in the field of ve- 7. (2017) Raspberry Pi 2: Which OS is best? | ele- hicular automation and will give researchers an immense ment14 | Raspberry Pi 2. scope to work on. 8. Vivek Kartha (2015) “Interfacing HCSR04 ultrasonic sen- Conclusion sor with raspberry Pi”. 9. https://elinux.org/Rpi_Camera_Module SoCs are like the much needed anecdote that provided the intermediatory solution to the struggle of choos- 10. https://www.bananarobotics.com/shop/How-to-use-the- L298N-Dual-H-Bridge-Motor-Driver ing between microcontrollers and computers. SoCs provide us with the required processing speeds, operating 11. Tarun Agarwal “Different type of microcontrollers are used system access, easy programmability like computers and in automobile applications”. reduced cost, size and easy hardware interaction like mi- 12. Lee Teschler (2017) “Centralized processing for autono- crocontrollers. In the field of ITS and automation we al- mous vehicles”. ways needed something which settled between comput- 13. Tim Holyoake (2016) “Benchmarking the raspberry pi 3”. ers and microcontrollers. SoCs seem to be tailor made 14. Paul Beckett (2015) “Some FPGA SoC applications and for those projects and applications. trends in advanced system on chip”. RMIT University, Mel- bourne, Australia. Acknowledgement 15. Rudolf Grave (2017) “The vehicle architecture of automat- The author appreciates support from the Department ed driving level 2/3”.

Citation: Mujtahid M, Sofi SA, Bashir T (2018) Comparative Advantages of System on Chips in Intelligent Traffic System. Int J Robot Eng 3:004