Study On The Fuzzy Control Of Automotive Automatic Parking

As the number of vehicles continue to rise and the scale of urban construction continue to expand, the difficult of parking in the traffic problems has become seriously. Automatic parking control system is the major auxiliary functions of intelligent vehicles. It is important to solve the problem of parking in narrow space. Also it will reduce traffic accident.This paper is the simulation study on the fuzzy control method of vehicles automatic parking. The research object was the kinematics vehicles model on Jetta cars parameters. According to the actual situation, this paper design fuzzy controller of the automatic parallel parking, automatic vertical parking, automatic slant parking.This paper mainly do the following studies :1. Establishing the kinematics vehicles model of front wheel steering2. Designing fuzzy controller of automatic vehicles parkingAccording to the practical background, this paper designed the fuzzy controller of Steering System for the automatic parallel parking, automatic vertical parking, automatic slant parking. Based on the analysis of simulation results, found that the selection of control input and control output, the selection of the preliminary parking location and the length of parking space, the width parking space have greater impact on the automatic parking control effect. Using the Fuzzy logic to design Steering Control System has completed the three control tasks. In the design of fuzzy control on vehicles automatic parallel parking, this paper has designed and analysed the target searching , fuzzy controller in the parkingprocess .Simulation results achieved satisfactory results. As the same designmentality, this paper has designed the fuzzy controller to the automatic verticalparking, automatic slant parking. Through simulation experiment discovered thatthe preliminary parking location and the length of parking space, the width parkingspace have greater impact on the automatic parking control effect. Because vehicleFront-wheel always has a lower speed moving in the whole parking process, thispaper used vehicles kinematic model instead of the vehicle dynamics model todesign the controller. And didn't consider the braking process and the startupprocess, this paper put focus on designing the control algorithm of shift controllerin the parking process, the simulation results is much more satisfactory.3. Using the genetic algorithms to optimize the fuzzy controllerBecause fuzzy controller is designed rely on experts' experience and testedconstantly, and that the design of the fuzzy controller is very time-consuming.Genetic algorithms based on the cost function for the objective function tocrossover and mutation. After several optimization of the generation to find anoptimal solution as the best one. Because the manual design of fuzzy logic controlsystem has blindness, empirical .This paper used genetic algorithm to design fuzzycontrol system. To make the membership function of fuzzy controller in theautomatic parallel parking better, found that the selection of cost function has agreater impact on control effect. The simulation results showed that the effect canbe improved, automatic vehicle can parked into the parking space which lengthis 1.4 times as Jetta car. The more important is that this paper has given the methodof the genetic algorithm how to optimiz fuzzy controller.Through the simulation research on vehicle automatic parking steering control system, founding the following conclusions : 1. Using three input variables and one output of a fuzzy control automatic parking system ,It can be completed automatic parallel parking, automatic vertical parking and automatic parking-control task better. Fuzzy Control System Design for parking space size and location of vehicles starting from have some adaptability and robustness.2. Using Genetic Algorithm to optimize the fuzzy controller that can improve the control effect. And the whole control system design of vehicle automatic parking is not rely on experts' experience and manual tests entirely any more, and design fuzzy control system automatically.