Signal Control Of Complex Intersection In Intelligent Network Environment

With the rapid growth of motor vehicles,drivers and the continuous development of the economy and society,traffic demand and traffic management pressures are increasing.The development of intelligent networked vehicle technology has provided a hardware foundation for more intelligent signal control,and also put forward new requirements for the next generation of signal control systems.In order to explore the boundary conditions and feasibility of intersection signal control problem,the characteristics of intersection signal control technology and complex intersection in intelligent networked environment was analyzed and summarized.From the perspective of system theory and cybernetics,intersection control problems in the environment was analyzed.Focus on the optimization of vehicle lane change and speed trajectory under fixed signal,overall control of vehicle and intersection under intelligent network environment,and overall control of intersection in hybrid intelligent network environment,specific research has been carried out from theoretical modeling,optimization algorithms and simulation experiment analysis.Firstly,considering the characteristics of the driving behavior of the intersection,the intersection entrance is divided into the reserve area,the lane change area and the isolation area.Under the minimum following distance constraint,the overall control strategy of the intersection is proposed.The passing vehicles are divided into the first car and the rear car,the optimal vehicle speed control model is established with the minimum exhaust emission and delay as the optimization target.The simulation platform was built with the Vissim software COM interface.The results show that compared with the traditional control method,when the flow intensity coefficient is 0.4-0.6,the proposed algorithm has the most significant improvement in the operating efficiency of the intersection,with the average speed increasing by 27%,the exhaust emission decreasing by 65%,and the delay decreasing by 67%.The low flow intensity,the correlation of lane change ratio and controllable area distance is weak;when the flow intensity is greater than 0.75,the correlation between lane change ratio and controllable zone distance is stronger.Secondly,a coordinated control model framework including signal timing optimization and vehicle trajectory optimization is established.A pass-right division model and a decision tree based on decision tree C4.5 are proposed.According to expert experience and simulation experiments,sample data is obtained for decision tree training.To prevent over-fitting,PEP pruning method is adopted.After testing,the decision tree is tested.The accuracy rate is higher than 95%.The signal timing optimization model is constructed,the high-dimensional solution space rolling optimization algorithm based on genetic algorithm and the enumerated lowdimensional solution space rolling optimization algorithm are established respectively.Finally,the overall optimization algorithm of traffic rights distribution,signal timing and vehicle trajectory is constructed.The simulation program was developed in python3.7,a lot experiments are carried out under conditions of changing control parameters.Results show that the proposed two types of control algorithms have the best improvement when the flow rate ratio is 0.23.The high-dimensional rolling optimization algorithm reduces delay by nearly 57.6%,and the lowdimensional rolling optimization algorithm reduces delay by nearly 44.8%.Finally,considering the safety factors of mixed conditions,the safety optimization model of driving safety and delay is established,the Pareto dominance of the problem is proved and the solving algorithm is constructed based on it.The different types of vehicle trajectory algorithms under mixed conditions are reconstructed,and the artificial vehicle prediction model under mixed conditions is proposed.Furthermore,the hybrid condition descending human phase,right-turning vehicle phase and trajectory control model are established.The high-dimensional and lowdimensional signal timing-time rolling optimization models were used respectively,and the overall solution algorithm was finally constructed.The experimental platform was developed with python3.7,a lot experiments are carried out under conditions of changing control parameters.Results show that the proposed algorithm improves the cross-over improvement with the penetration rate of intelligent network vehicle,and the optimal gain point is 50%.The adaptability of the overall control algorithm to real-time changes in traffic demand has been experimentally verified.When the flow rate ratio is less than 0.45,the algorithm has the best control effect.