Study Of Dynamic Area Coordination Control Of Urban Traffic Based On FCD

In recent years, total motor vehicles rapidly increase in China and urban traffic congestion has become more and more serious. It is impossible to solve urban traffic problems only reliance on the construction of transportation infrastructure. Intelligent Traffic Control is developing trend of modern traffic. There are some difficulties in the traffic control systems such as control model, real-time traffic information acquisition. This dissertation study urban area coordination control method under the above conditions.The main achievements are as follows:Firstly, to solve traffic data acquisition problem, this dissertation studies method which calculates traffic flow by lane average speed. Floating car data technique makes it possible acquisition lane average speed in urban road network. Through analyzing FCD characteristics, we propose a FCD map-matching algorithm based on local path searching. The maximal possible distance is taken as heuristic information to reduce search area. This algorithm utilizes the information of previous matched GPS position, and generates the rectangle confidence area centered the matching GPS data to determine the candidate match paths. The direction of FCD is combined with"point-curve"distance to accomplish the map-matching. The experimental results demonstrate this proposed algorithm has much higher accuracy rate than the typical navigation map-matching.Second, through analyzing the classic Speed-flow models under different traffic conditions, logarithmic model is chosen to represent the correlations between speed and flow. The relations among the adjacent intersections are addresses. On these bases, we propose a urban area coordination control algorithm based on Q-learning. The proposed algorithm is validated with microscopic traffic simulation software, and the experimental results demonstrate this proposed algorithm has better effective to relieve area traffic jams.