Freeway On-ramp Control Strategy Of Single Point

As traffic flow increasing, congestion on the highway and blocking phenomenon occurs after another, greatly reducing the operating efficiency and highway safety, energy consumption and environmental pollution make the problem worse, the study of a reasonable and feasible control methods is an urgent need. On-ramp control approach is an effective measure to solve this problem, by controlling the highway on-ramp flow so that traffic flow on the main highway running in an desired state, this can effectively prevent the occurrence of the congestion problems. The characteristics of highway traffic system is analyzed, and the contents and conclusions of the study are as follows:After analyzing the characteristics of traffic flow on the highway and the relationship between the various parameters we know that highway system is a strongly nonlinear system. In the fact that the neural network has good nonlinear approximation properties and anti-jamming capability, the BP neural network and PID control algorithm are introduced to the freeway on-ramp control, by adjusting the on-ramp rate to maintain the desired traffic density on the main highway. Theoretically rigorous analysis proves the convergence of this control algorithm, simulation results show the effectiveness. The structure of the controller algorithm is simple and easy to achieve, for this highly nonlinear system has a good control effect.Based on the characteristics of BP algorithm and simulation results of its actual meaning, an approach called RBF neural network combined PID controller has been proposed, and the comparison of the results of classical ALINEA are given. Because of the RBF algorithm overcomes the inadequacies of BP algorithm and classical ALINEA control strategy, the stability of the system become better and the anti-perturbation ability become stronger under the same case. Theoretically rigorous analysis demonstrates that the convergence of this method and simulation results show the superiority.The nonlinear approximation ability of neural networks and its self-learning ability make the method has a good adaptability when system suffering the perturbation. The upstream traffic is taken into account while the main highway traffic flow density is ensured that always at a desired density of its neighborhood. Studies have indicated that combining the neural network PID control technology with the single point of highway on-ramp control is helpful to ease highway congestion and control mainline of the density fluctuations.