Traffic Signal Predictive Control Based On Cloud Platform

With the development of the city and the improvement of living standard, car ownership is growing rapidly, which brings tremendous pressure to the urban traffic system. At the same time, traffic congestion, traffic accidents and environmental pollution appear frequently, which has seriously affected the efficiency of people traveling and quality of life. When the road capacity can’t meet the demand of traffic, there is no way to get treatment timely, so the traffic congestion occurs. With the traffic demand increasing, road capacity can’t meet its needs. When the traffic demand increases to a certain extent, traffic congestion will occur at the road intersection first. The traffic congestion may travel around or spread, which forms a "Domino Effect". That is the reason of a wide range of traffic congestion. The scientific and reasonable design of an intersection signal, can effectively increase the intersection traffic volume, so as to avoid the occurrence of congestion. Of course, it is not enough to control an intersection, the regional road network coordination control can ensure smooth traffic. Therefore, it is important to explore the regional traffic signal control system.In this paper, according to the theory of predictive control, the strategy of predictive control of traffic signals is put forward based on the non-analytical model to solve the optimization model. These links are inseparable from the efficient handling of data information. Since the date of birth, cloud computing is paid attention by the researchers in various fields. The ability of quick service, fast processing capacity and elastic computing ability of cloud computing provides the opportunity for its development in the field of traffic information processing. Therefore, on the basis of using the cloud computing technology to analyze and deal with vast amounts of traffic data intelligently, studying the traffic signal predictive control based on cloud computing provides the theoretical basis and the technical support for the governance of traffic bottlenecks and the reference of development and application of cloud computing in the field of transportation.The following missions have been finished in this article:First, this article builds up a traffic signal predictive control strategy and a traffic flow simulation model has been built up, which is based on the research of predictive control mechanism and cellular automaton theory. Second, the process of creating a traffic flow simulation model and the realization process of the parallel genetic algorithm has been established. The parallel genetic algorithm is realized on a Hadoop Cloud computing platform, and the model is programmed based on MapReduce. By using parallel genetic algorithm to solve the optimization model, which synchronously optimizes the signal control parameters in a region, the signal timing schemes has been optimized in this model.Third, the superiority of traffic signal control based on predictive control with comparison of the traffic signal predictive control, timing control and inducing control in multiple times of simulations has been approved. The simulation model is programmed under the C++ Builder and Java environment. And the main functions of this model are simulating the vehicle operation in the road network, optimizing the timing of intersection signals and realizing the program. It has been also approved that the traffic signal predictive control can effectively improve the traffic capacity of a road network by simulating the minimum number of waiting passing vehicles in 4 crossroads selected from road networks.The method of using the cloud computing platform to establish the traffic simulation has been clearly demonstrated in this article. This method has successfully realized the combination of simulation technology and optimization technology, and the combination of predictive control and traffic signal control has obtained the expected results. Finally, the direction for future research is provided.