Study And Design Of A New Intelligent Traffic Signal Controller

With rapid urbanization process, urban traffic congestion is becoming increasingly serious, and grade intersection is the main bottleneck of traffic jam. Because it can effectively alleviate increasing congestion, intelligent transportation system has been the development direction of modern traffic. Intelligent traffic signal control is one of the important research fields in intelligent transportation system. Traffic signal controller of intersection is the infrastructure for intelligent control of traffic signal. The technical level of the signal controller directly determines its role in the traffic signal control system. Thus, it strongly influences effectiveness of urban road traffic management.In this dissertation, current development status of traffic signal controller is discussed in detail. In order to adapt trend of intelligent transportation, a new intelligent traffic signal controller is presented. Its hardware structure, reliability, software design and single-intersection signal control method are studied and analyzed. The main work of this dissertation is as follows:1. A new design way of intelligent traffic signal controller is presented, and modularization method is used to develop the hardware system of signal controller. Aiming at the shortcomings such as functionality and expansibility of the current signal controller, a modified design scheme is given. It involves in the use of high-performance ARM9 chip as the processor, high-capacity flash memory chips, adding the network interface, using CPLD to expand I/O pin, feedback circuit of the lights state, the use of touch-screen interface as the input and output. The detailed description of the hardware design is given in dissertation.2. Reliability and security of the traffic signal controller are necessary. Reliability scheme of signal controller, which includes the circuit system, fault self-diagnosis and lightningproof, is developed. Fault self-diagnosis meets the requirement of safety. Two-level detection method using software and hardware ensures the safety of light colors output.3. After software development environment is established, the basic hardware device drivers and the main control program are implemented, and the communication protocol between traffic control center and signal controller is defined. In order to meet the requirement of humanization operation, graphical user interface is designed and implemented using Qt.4. Signal control of single intersection is the foundation of large-scale traffic signal control. Firstly, two basic control ways—fixed-time control, vehicle actuated control—are analyzed. Then, reinforcement learning is studied. An optimal control algorithm of single intersection based on Dyna-Q is proposed. The simulation results indicate this algorithm is effective.The traffic signal controller developed by us meets the requirements of network, intellectualization and humanization. It accords with the trend of traffic signal controller and provides a useful reference for further research.