| As an essential part of alleviating urban traffic congestion,effective traffic control primarily focuses on addressing the urban road traffic problems including the complexity of road networks,the scale of traffic flow,the regionalization of congestion and the confusion of control,etc.To target solving the above issues,this dissertation carries out some research and applications on control technologies for urban regional roads.The main innovations of this dissertation are as follows:(1)A sub-division algorithm of the urban regional road traffic network based on the distribution of the congestion index is proposed.With this algorithm,the vehicle trajectory data could be converted into a trending urban road traffic flow congestion index,and the division of road network sub-areas could be transformed into the graph division.Besides,the highly congested nodes could be divided into the same subset as their adjacent roads and intersections with the same congestion level.The experimental results indicate that the proposed algorithm has the characteristics of fine granularity,high cohesion,and low coupling.In the case of the adjacency list used as the graph storage structure,the time complexity of our proposed algorithm is between O(nlogn)and O(n~2).(2)An urban regional road traffic coordination control algorithm based on road network subdivision is formulated.In which,the coordinated control of each sub-district of the road network is regarded as a multi-objective function,that aims to realize the goal of maximizing the regional road traffic output capacity and minimizing the traffic delay.In this way,the result could be optimized with the genetic algorithm.The traffic output capacity of proposed algorithm presented in this dissertation is 1.35 times of the single-point timing control and 1.12 times of the coordinated control without network subdivision.Furthermore,our algorithm has an obviously better congestion mitigation advantage in lightly congested road conditions than in moderately and heavily congested road conditions.(3)A dynamic path planning algorithm for urban road traffic based on our improved ant colony algorithm is suggested.In which,the road condition factor could be jointly established according to the main attributes of urban roads,such as the length of the urban road,the number of lanes,the incoming traffic flow,the outgoing traffic flow condition factor,etc.The created factor could be exploited to replace the distance parameters in the particle swarm algorithm and ant colony algorithm,and the parameters of the ant colony algorithm could be optimized with the particle swarm algorithm,which makes the improved ant colony algorithm more suitable for dynamic path planning,especially under traffic congestion.The experimental results demonstrate that the proposed dynamic path planning algorithm can improve the effectiveness and accuracy of path planning,which can reduce the congestion rate by 10～14%compared with the ant colony algorithm based on distance parameter.(4)An urban area road traffic simulation system based on distributed computing is designed and implemented.Our main task is in detail to design and apply the simulation system from the aspects of functional composition,collaboration relationships,and task execution.The experimental test results show that the proposed distributed computing-based road traffic simulation system can not only support the urban regional road traffic simulation of different vehicle scales in terms of simulation scale,but also feature good scalability and high efficiency.The main work of this dissertation covers theoretical research and technical applications.In detail,the techniques of sub-zone division,coordinated control,and dynamic path planning can realize traffic control and congestion mitigation on urban roads.The prediction,verification,and analysis of traffic control are implemented via simulation systems.Our research results can be applied in many fields,such as urban management and traffic engineering. |
PDF Full Text Request