Research On Geometry Algorithm And Fuzzy Logic For Automatic Garage Parking Of Sedan

Automatic parking system has been regarded as next hot product of automobile industry. It can alleviate driver's physical and brain's load and reduce traffic accident. At the beginning of a project, simulation and analysis via math model and control strategy can not only adjust design scheme flexibly and optimize design parameters, but also reduce research expense and shorten develop period. Automatic garage parking, which consists of port detection and steering control, is one of two working models of automatic parking system and steering control is the key point of the model. This paper focuses on steering control algorithm and the premise is that the vehicle can detect effective port and localize itself accurately.This paper tries to find easy and utility steering control algorithm from variable control methods. For the topic, research work and the content are as followed:In Chapter One, it introduces the developing actuality of automatic parking system and algorithm of back-up, the significance of automatic parking, the research and application in vehicle industry of fuzzy algorithm. In the end the main content is showed.In Chapter Two, the theory analysis of back-up is carried on. Firstly, it describes the whole process of garage parking and confirms the parameters of vehicle. Secondly, it confirms the reference garage via investigation of garages in the Changchun city and reference to the drive license examination in Jilin Province. Meanwhile the reasonable back-up speed is made according to the assessment of experienced driver. Finally, it deduces the kinematics equation when vehicle is with low speed. The relationship between the steering angle of wheel and the steering angle of middle point of front axis is analyzed. Then Simulink model is built.The main content of Chapter Three is the algorithm which is deduced by geometry. Ordinary driver parks the sedan at a place according to the relative position between the sedan and the garage at first when he backs up. Then he turns the wheel with the maximum angle .He revises the angle gradually till the sedan parks at an ideal position .In the chapter, feasible region is defined and found at first. It confirms reference parking region .It is apparently that the sedan can not be parked at the feasible region accurately each time .So an estimated algorithm is brought out. The simulation result of the estimated algorithm shows that the algorithm can resolve the garage parking problem when the vehicle is with different initialized condition .In the end, a shortest trajectory is calculated according to the REEDS's algorithm and the two algorithms are compared.The main aim of Chapter Four is to research algorithm using fuzzy logical. The design of fuzzy logical system is always laid on experience and data. For the reason data collecting experiment is made. Distance and yaw rate are collected when the vehicle with different initial positions and orientation angles. The relationship between X coordinator, Y coordinator, orientation angle and time can be obtained via the processing of collected data and initial pose. Good data is used to determine the fuzzy logical system through the form-checking method and observational method. The system is simulated in Simulink. Then anfisedit, which is a tool box of MATLAB, is used to build adaptive-network-based fuzzy inference system of garage parking. At last the four algorithms above are compared.The Chapter Five is aim to validate the model and algorithms above using Carsim software and vehicle experiment. Vehicle is complicated system .The mutual affection between kinematics and dynamics is one reason of the complication. Carsim is good software at simulation for whole sedan because it can accurately output status variables during the responding period. The steering angle which is outputted by the algorithm above is used as an input of Carsim software and the outputted trajectory indicates the sedan can park to the garage successfully. The knowledge about ultrasonic sensors is not mastered for the research team at the moment. In the case, the experiment is based on the steering angle and distance. Several results of experiments show the algorithms in the article are effective.Chapter Six is the summarization and further work. The research in the article, shortage and further work are stated.This article, which has built the foundation for the platform of automatic parking system, has done some research on the steering control algorithm.