Study On Automatic Parking Control Strategy For Parallel Parking Scenarios

The automatic parking technology has important practical significance for alleviating the parking problem,reducing the driver's parking pressure,improving the comfort of travel,and saving the economic cost of driving.Accurate and effective automatic parking guidance and control strategy can direct the vehicle smoothly and safely into the designated berth.The current mainstream method of automatic parking is generally planning a parking trajectory vehicle can generally follow based on the first special curve(such as ?-spline curve and polynomial curve,etc.),and then makes vehicle roughly to track the trajectory planning by adding a certain amount of feedback control.However,this method is difficult to be totally considering constraints of vehicle dynamics in trajectory planning stage,and when the trajectory planning phase errors,it is difficult to through the feedback control to make up for this mistake.In this paper,for a parallel parking scene with more complex environment and higher requirements for driver technology,a kind of "direct guidance control" parking method and control strategy referring to manual parking experience was proposed.To avoid the disadvantages of direct trajectory planning,the paper studied the parking"initial position-final position" and corresponding control instruction data,established the relationship between the two and store it into the system.When vehicles are faced with a new parallel parking scenario,the system can directly"search" the current scenario corresponding control instructions,thus use this instruction fast guide vehicles for parking in the absence of path planning.In order to obtain the "initial position-final position" and corresponding control instruction data in the parking process,on the premise of constructing reasonable parking scene and establishing proper vehicle dynamics model,the paper put forward a method of parking trajectory planning based on the experience of artificial parking:on the basis of the analysis of the current scene parking feasibility,the evaluation criteria of optimal starting position and optimal control input combination are established,and according to the control instructions calculated based on the standard to guide the vehicle reverse movement.In addition,based on the robustness requirement of parking system,the paper analyzes the factors that cause the deviation of actual parking trajectory and planning trajectory.The linear quadratic optimal controller and the real-time negative feedback controller based on the system state are designed to reduce the above deviation.