Application Research On Traffic Control Based On Data-driven Control Approaches

Along with the rapid growth of motor vehicle quantities, traffic jam phenomenon is more and more common. How to reasonably use the existing traffic facilities and improve the level of management and control become an effective way to solve the traffic problems.This dissertation focuses on applications of data-driven control approaches in queuing length balancing control of urban oversaturated isolated intersection, relative occupancy balancing control of arterial and on-ramp&off-ramp control of freeway. The main research contents and key innovations are summarized as follows:1. Based on store-and-forward modeling, apply the queuing length balancing control to isolated intersection using data-driven control methods such as iterative learning control (ILC) when traffic demand satisfies the repetitiveness or non-repetitiveness situation. The convergences of these improved algorithms are proved respectively. The queue of each phase tends to be equal. The effectiveness is verified via PARAMICS platform.2. In order to reduce the on-line amount of calculation, the queuing length balancing control of isolated intersection via multi-parametric programming is given. The definition of performance index of queuing length balancing is proposed and the signal timing is solved off-line. The the signal timing strategy is obtained on-line according to the convex area of queuing length parameter vector. The effectiveness is also verified via PARAMICS platform.3. The definition of critical synchronization and critical relative occupancy is proposed. Based on the basic idea of ramp control, the relative occupancy balancing control of main trunk road is applied to maximize the output of control system and avoid wasted green time, spillover, gridlock via adjusting the signal timing of upstream and downstream intersections. The effectiveness is also verified via PARAMICS platform.4. Based on the detection data of the actual operation of freeway, data-driven control methods such as iterative feedback tuning (IFT), virtual reference feedback tuning (VRFT) are firstly applied to tune the ALINEA controller parameter and also verified via PARAMICS platform. Active disturbance rejection control (ADRC) is applied to freeway ramp metering and the effectiveness is also verified via PARAMICS platform. Vehicles of off-ramp can not travel off the main road rapidly. Coordination control strategy of off-ramp and downstream intersection of side road is designed and also verified via simulations.