Research On Automatic Parking Path Planning Under Parallel Parking Scenario

With the Xiaokang approaching its final stage and the import and implementation of various State Economic Measures,national economic development has been accelerated and people’s living standards have been significantly improved,cars are no longer luxury goods but enter national life as ordinary consumer goods.The number of cars has been increasing year by year,while the space of parking spaces has become more and more compact.Parking difficulties have also become a common social phenomenon,the intelligent automatic parking system brings a good solution to the problem that parking is difficult for people.In recent years,the development of driverless technology has provided technical support for the realization of automatic parking system,which has been regarded as a research hotspot by many mainframes and research institutions at home and abroad.The main components of automatic parking can be divided into perception,decision-making and executive layer.In this paper,the path planning in the decision-making layer is studied.Firstly,the application of ultrasonic sensor in the parking space detection of automatic parking system is introduced,and its ranging principle is briefly analyzed,to determine the parallel parking detection method,through the perception layer in theory to lay the foundation for research.Secondly,the motion of the vehicle at low speed is studied,the vehicle turning angle is simplified through Wilhelm Ackermann’s turning principle,and the vehicle kinematics model is established by integrating the kinematics model and dynamics model,based on the highprecision vehicle simulation software VEDYNA,the model is verified to meet the requirement of controller design in predictive time domain.Thirdly,this paper planed the parking path by analyzing the two-arc tangent straight line,and optimized the path by analyzing the curvature continuity of the planned path and the EPS steering system load,but only the EPS system load problem is solved.In order to solve the continuous curvature problem of the planned path,the final optimized path planning is obtained by planning the path of two transition curves,three path planning schemes are simulated using Matlab / Simulink software platform.Finally,the model predictive control is chosen as the path-following controller of the system by analyzing the characteristics of the model predictive control(MPC)for predicting the future motion characteristics in the limited range of the parking scene,according to the three paths designed,the simulation results of tracking error verify the effectiveness of the optimized path tracking method.