Intelligent traffic controller

Abstract

INTELLIGENT TRAFFIC CONTROLLERBy DESAI MA YURKUMAR JA YESHBHAI MASTER OF SCIENCE IN ELECTRICAL ENGINEERINGAutomotive technologies are gaining ground in modem road traffic-control systems, since the number of road vehicles and passengers is rapidly growing. There is a perpetual need for safety-critical traffic automation, and traffic engineering makes the dynamic or static analysis and the synthesis of automotive vehicle technologies possible. The main goal of engineering is the planning and management of traffic systems. The project supports the development of reliable and optimal control structures for urban traffic and for motorway systems.The intelligent and cooperative set-up of actuation and its linkage to the central control system is vital for avoiding traffic jams and accidents. Moreover, environmental costs (eg pollution) can be decreased. One aspect of the project aims at developing a traffic control algorithm for future technology. The design of the traffic control system can be evaluated in two steps -synthesis and analysis. Several models and multiple control strategies exist, and engineers must decide between them using a priori knowledge of the real system.Previously collected information can help to choose the appropriate model, parameters, measurement and control methodologies to create the optimal solution. In many cases, control-related variables are almost inaccessible for design unless estimation techniques are applied. In a situation like this, the approximation, computer-based estimation of the variables could be useful. Traffic simulations can be classified in several ways, including the division between microscopic and macroscopic, and between continuous and discrete time approaches. Several state variables, derived from the description of the dynamic system, can be used for operational and planning aspects. A newly emerged area is demand estimation through microscopic traffic modeling. The dynamic aspect of traffic simulation requires previously measured or estimated volumes of traffic.Since the measurement of certain variables in the dynamic description is rather costly, one tries to estimate them. For instance, the observation of constantly varying turning rates at a simple intersection is fairly costly. However, the number of turning vehicles could be applied to traffic light harmonization, or generally speaking to traffic light control. This automated traffic signal controller can be made by suitably programming GAL device. The controller assumes equal traffic density on all the roads. In most automated traffic signals the free left-tum condition is provided throughout the entire signal period, which poses difficulties to the pedestrians in crossing the road, especially vii viii

Smart Cities Trafic Light Management Systems Review: Models and Approaches

“…With proper integration of STMLS, most congestion problems can be solved. Indeed, multiple solutions have been proposed in the STMLS specialized literature, either for a single intersection [1,2,8,10,12,[16][17][18] or for a set of intersections [11-13, 3, 5, 19] and that according to different approaches [1][2][3][4][5][6][7][8][9][10][11].…”

“…This problem has further attracted the attention of researchers with respect to the emergence of the intelligent city model. Although the design of STMLSs is based on different approaches such as the use of sensors [1,16] optimization methods [5,8,10,11] and the combination of the two previous ones [2,3,12], these STMLSs are still subject to improvements to comply with the requirements of smart cities.…”

“…They are generally classified into two principal categories: models for a unique intersection and models for a group of intersections. For both models, the approaches used were: heuristic approaches [8,10], sensor-based approaches [1,[16][17][18] and hybrid approaches [2,12] that combine the first two ones.…”

“…To reach acceptable performance in real-world applications and with any solution, the goal of STMLS is to minimise the waiting time of cars and the length of queues and to maximize the total flow of vehicle trafic on roads by using different optimization methods. These include dynamic programming [1], artificial neural networks [2], mathematical models [3], bee colony optimisation [4], multi-objective diversity-based evolutionary genetic algorithms [5], multi-agent [6], fuzzy logic [7], reinforcement learning [8], game theory [9], and Petri nets [10]. Sometimes, a mixture of two techniques such as: multi-agent and deep learning [11] are used.…”

“…Very few solutions take up the enormous amount of data available with the most modern wireless captors. In this context, some STLMS systems use only optimization approach [10-13, 5, 8] or data collected by the sensors [14][15][16][17]1] while others use hybrid approach combining sensors and optimization techniques [2,3,12]. Therefore, the best way to fully reaching the objective of STLMS in a smart city is not only to analyze the different approach models and their collecting data mechanisms but also to keep in mind that an STLMS is just an embedded system located at only one intersection.…”