The Research On Automatic Parallel Parking Path Planning

With the continuous increase of car ownership and urban land price,the parking spaces are smaller and smaller and the parking in narrow spaces may bring many inconveniences to drivers.For example,the parking time too long hampers traffic,and collisions,scrapings and others occur during parking.Automatic parking system has aroused great concern in solving the parking problem.The automatic parking system can reduce the operational effort in parking for operators,reduce the occurrence of parking accidents effectively and improve the safety and comfort level in parking.In this paper,a car kinematic model is established to analyze the steering characteristics of cars at low speed according to Ackerman steering principle,and an automatic parallel parking path tangent to arcs is planed based on it.A reasonable parking endpoint area is designed in a parking area by finding the initial area where the parking can be completed according to the limit steering characteristic of cars and estimating the shortest parking length.The feasibility of this method is verified by planning the first parking path of endpoint regionalization and the secondary parking path of it according to the size of parking space,analyzing possible collisions during parking,establishing constraint functions and establishing a simulation model for automatic parallel parking in Matlab.The comparison between the secondary translational parking of endpoint regionalization and the secondary parallel parking of it in narrow spaces shows that collision-free parking can be achieved with the two methods,while the secondary parallel parking is more desirable.In order to improve the efficiency of secondary parallel parking with endpoint regionalization and realize the shortest parking path,a path function for secondary parallel parking of endpoint regionalization with constraints is established and the function is optimized with genetic algorithm.A simulation model for secondary parallel automatic parking based on genetic algorithm is established in Matlab and it shows that cars can be parked within the designed endpoint regionalization without collision correctly and the shortest path can be realized by optimizing the parking path function through genetic algorithm.In this paper,an overall structural framework is designed for the automatic parallel parking control system,and a detailed design experiment is finished for the detection system.The experimental results show that the design plan can effectively detect the distance between a vehicle and an obstacle.