Application of CACMAC-PID Composite Control to Highway Density Control via Ramp Metering

摘要

To solve the traffic congestion problem in the highway mainlines and to improve ramp metering effect, we propose a road density control approach based on the traffic model and compound control. This compound control is a combination of proportional-integral-derivative (PID) controller and credit assignment cerebellar model articulation controller (CACMAC). Firstly, a cell transmission model (CTM) is formulated, which is an advanced technology means for simulating highway dynamic traffic flow. Then, CMAC (cerebellar model articulation controller)-PID composite control and CACMAC-PID composite control are investigated. CACMAC uses a credit assignment CMAC learning algorithm to raise learning efficiency and to attain the expected accuracy quickly. Highway density control is realized by using feedforward control and nonlinear feedback control, as well as by integrating the CTM. Finally, system simulation is carried out by two different traffic situations. The results show that CACMAC-PID composite control has superior density tracking effect and stronger interference suppression capability compare to CMAC-PID control. The composite control can eliminate highway traffic congestion quickly. Obviously, the proposed compound control approach is especially suitable for highway density control.