Research On Data Driven Control Method For Expressway Entrance Ram

As the main skeleton of the urban transportation system,the freeway carries a large number of motorized travel tasks.However,with the rapid increase in urban motor vehicle ownership,the freeway transportation systems in major cities have poor operational efficiency and often encounter problems such as traffic congestion.The on-ramp is the most effective control method to regulate the efficient operation of the freeway,that is,by regulating the traffic flow of the on-ramp entering the freeway,to ensure that the traffic flow it carries operates in the best service state.In view of the above analysis,this paper designs an on-ramp regulation scheme based on the data-driven control method for the macro traffic flow system of the freeway.The design of the data-driven controller does not need to establish a complex freeway traffic system model,but only uses the input/output(I/O)data design of the system.The main works and innovations of this paper are summarized as follows:1.In view of the complex,time-varying and difficult modeling characteristics of the freeway traffic system,a freeway on-ramp regulation scheme based on model-free adaptive predictive control(MFAPC)is designed,and its stability is analyzed.The advantage of MFAPC controller is that it is a typical data-driven controller,and the system model does not participate in the controller design.It only utilizes the I/O data of the system,eliminating the dependency of control system design on the controlled object model;at the same time,MFAPC introduces future I/O signals,which have stronger robustness and wider applicability.Further,in view of the obvious repeatability of the freeway traffic system,an open-loop iterative learning control is introduced based on the MFAPC scheme,and an MFAPC freeway ramp regulation scheme with an iterative learning outer loop is proposed,and the stability of the scheme is analyzed.Compared to the MFAPC scheme,this scheme can utilize iterative learning feedforward controllers to compensate for repetitive disturbances in the system,and can achieve complete tracking control of the system.It is worth noting that predictive controller and learning controller can work independently or jointly.Finally,through traffic flow simulation experiments,it can be concluded that the proposed method in this paper reduces the total travel time of the freeway traffic flow by about 0.33%,extends the travel distance by about 0.06%,and increases the average speed of the traffic flow by about 0.18% compared to the ALINEA and other on ramp adjustment schemes.2.A new type of data driven iterative learning control(DDILC)on freeway ramp adjustment scheme is proposed for the freeway traffic system with repetitive characteristics but the initial state is not strictly repetitive,and its convergence is analyzed.Compared with traditional iterative learning control scheme,DDILC scheme relaxes the same initial conditions,and the learning gain can be adjusted iteratively,avoiding the blindness of learning gain setting.Further,considering that the system may have external variable disturbances or internal uncertainties,a robust data-driven iterative learning control(RDDILC)on-ramp regulation scheme with disturbance compensation is designed,and the convergence analysis of the scheme in the sense of maximum norm is provided.Theoretical research shows that the RDDILC scheme is not strictly restricted by the same initial conditions in the iterative process,but also relaxes the repeated external disturbances in the iterative process.Finally,through traffic flow simulation experiments,it can be concluded that the proposed method in this article reduces the total travel time of the freeway traffic flow by about 3.25%,extends the travel distance by about 0.25%,and increases the average speed of the traffic flow by about 0.38% compared to the ALINEA and other on ramp adjustment schemes.