A Study On Path Planning And Tracking Control Strategy Of Parallel Automatic Parking Based On Ultrasonic Ranging

In recent years,with the continuous increase of domestic car ownership,parking space is becoming smaller and smaller.In addition to the uneven driver technology,more and more traffic accidents are caused by parking.The automatic parking system can park the car into the parking space efficiently and safely,reduce the parking burden of drivers and the probability of accidents.The parking space detection,path planning and tracking control of parallel parking are focused on this thesis.Firstly,the Freescale intelligent car is equipped with ultrasonic ranging module to realize the parking space detection function.Aiming at the influence of environmental temperature on the speed of ultrasonic transmission,the temperature compensation program is designed to reduce the measurement error.Through the design of comparative experiments,it is proved that the use of ultrasonic detection of parking space is feasible.Secondly,based on Ackerman steering geometry principle,the kinematics model of low-speed parking is established.Combined with vehicle structural parameters,the position coordinates and trajectory equations of key points are derived.By comparing the simulation results of Carsim and MATLAB / Simulink vehicle rear axle center trajectory,the established kinematics model of low-speed parking is proved to be valid.Then,the minimum parking space required for parking is calculated by using the critical condition that the vehicle drives out of the parking space with the minimum turning radius.Aiming at the problem of large angle turning at the common tangent point of the arc tangent planning parking path,the circular arc straight arc parking path is designed,the feasible starting area of parking is analyzed and calculated,and multiple points are selected on the boundary of the starting area for simulation.The result shows that the problem of large angle steering in situ is solved;aiming at the curvature discontinuity existing in the transition between arc lines,the cubic third-order B-spline curve is proposed to optimize the parking path,and a non-linear constrained parallel parking path optimization function is established based on the impact environment constraints,vehicle self-constraints and parking requirements.Through multi group simulation,the optimized parking path curve is proved the rate is continuous.Finally,based on the model predictive control theory,the low-speed parking kinematics model is linearized and the objective function is determined.Under the constraint conditions,combined with the optimized parking path,the model predictive parking path tracking controller based on the given path is established.The sine function is tracked by Simulink simulation.The result shows that the designed path tracking controller is reliable,and can be used for parking path tracking.Joint simulation platform of Carsim and MATLAB/Simulink is established to simulate and verify the tracking effect of the path tracker designed in this thesis,the result shows that the parking path tracking controller based on model predictive control has good tracking performance.In order to verify the tracking effect of the designed tracking controller,the tracking effect of the pure tracking algorithm on the given parking path is compared horizontally.Compared the tracking error of the two algorithms,the result shows that the parking path tracking controller based on model predictive control has better tracking effect and smaller error,which meets the design requirements.