Car brings a great convenience to mankind. However, the difficulty of driving a car,especially in parking makes many people feeling not so good. An effective solution is to usethe automatic parking system replacing man-made parking. As the most difficult part of theautomatic parking system, automatic parking control strategy is a worthy problem.In this article, we solve the parallel parking problem by a non-linear path-based feed-back control, through the no-skidding vehicle model established on the plane. And we makethree mainly studies.Firstly, we achieve the nonholonomic constraint system of car in plane, using Acker-mann turning principle and wheel sliding restrictions. And we get the relation of the vehicleturning radius and the speed and steering angle of vehicle. On this basis, we get the mini-mum parking space length and the minimum road width required, by two tangent circle.Secondly, we establish the crash safety region in parking by vehicle collision con-straints and using the vehicle parameters and road parameters. In the area of crash safety,we build the function of the curvature of the parallel parking trajectory using the vehiclestarting vertical axis and the maximum steering angular speed constraints. After this, we getthe parallel parking trajectory by the curve fundamental theorem.Finally, we get the vehicle dynamic system in curvilinear coordinates by vehiclenonholonomic constraints kinetic equation using nonlinear transformation. By nonlineartransformation, the kinetic equation can be converted to a nonlinear control system whichis topological homeomorphism with linear control system. Above all, we get a stabilizednonlinear feedback control and use the Lyapunov function to prove it. To demonstratethe effect of feedback control of nonlinear path control, we use Matlab to get the vehicleparallel parking simulation. |