Increasingly tight parking spaces in urban, a driver should have the necessary skills making the car into a small space. But for most of the drivers, parking is a very complicated and painful things, even if general parallel parking spaces and back-up storage, etc. Many scholars and experts research into automatic parking system, which could reduce the burden of the driver's and provide a relaxing and comfortable environment. Automatic parking system can make a car park into parking space automatically and accurately.To make a car of domestic A0 Class as object, this paper researches the vehicle steering control algorithm, used to horizontal line-up parking and on-street docked parking which are called parallel parking.First of all, this paper introduces prospects and situation of home and abroad simply and describes the topics background of academic and application, and gives the main content.Secondly, this paper analyzes the factors of reversing speed and estimates the scope of reversing speed. Analysis of the vehicle kinematic characteristics during reversing, because the reversing is a low speed process, suppose the tires don't occur sideslip during the back-up course, thereby this paper builds a back-up kinematic model of vehicle. Analysis of reverse model, the reversing trajectory of rear wheel depends on wheelbase, tread and he Ackerman steering angle, has nothing with reversing velocity. This paper has done a set of low-speed reverse circular motion test of the target vehicle to verify rationality of the back-up kinematic model. According to the Ackerman steering principle, this paper analyzes the steering characteristic and transmission ratio experiment of the car of domestic A0 Class at a low speed reversing and build a model of steering system. The connection of wheel steering angle and Ackerman steering angle can be regarded as linear.Next,based on fuzzy logic control theory, analysis of the trajectory of parallel parking as well as the possibility of reversing collision, summary steering wheel rotation law of the skilled driver, this paper designed a simple fuzzy controller, which has the ratio of horizontal-vertical coordinates of ground point of rear tires and length and width of parking spaces, heading angle as control variables. The fuzzy rules is acquired simply from subjective driving laws of parallel parking, although the car under the controller can complement parallel parking better and the controller have good adaptability of the size of parking spaces. The controller doesn't consider objective condition, so its application isn't good, the rotating speed of steering wheel angle is too fast, and the back-up starting area isn't determined.In this paper, two main approaches-path planning and adaptive neural network system have been used to improve p reviouscontrol strategy.In view of the vehicle kinematic characteristics when reversing, this paper preliminarily estimates the size of parallel parking space which can make the domestic A0 Class car complete parallel parking based on limit turning radius of the car, and plans a trajectory which consists of 2-arcs-tangent, Analysis of the possibility collision between vehicle body and parallel parking space, between body and some barrier in the 2-dimensional coordinate system of parallel parking which has been built in the paper, based on reversing trajectory of 2-arcs-tangent curve, this paper acquires minimum size of parking spaces, minimum size is within the range of the space size which has been acquired from limit turning radius, so the minimum size is reasonable. According to collision-free, a number of trajectory has been identified and reversing strategy is determined. According to the 2-arcs-tangent reversing path and the range of trajectory, the starting area is been determined, from where the car can complete the parking course. According to collision of body peak when the car has been reversing with various speeds of the limits permitted, the parking space is modified to solve the problem of velocity fluctuation in the back-up course. The rate of steering wheel angle has a big change when tracking the two-arcs-tangent trajectory, which make the complement of parallel parking more difficult. The arctangent continuous function is selected to fit the original trajectory, which make the rate of steering wheel angle continuously. At the same time to reduce approximation errors as a result of the curve characteristics, steering wheel angle in the later of parking course is improved, which solve the problem of failure parking because of straighten body unsuccessfully, when the car is tracking arctangent trajectory in the later of parking.In order to acquaint reversing features and operation feature of parallel parking better, this paper has collected two sets of back-up experimental data with data acquisition system which has been developed independently. The first set of experimental data is acquired from the uniform circular reversing motion; in this experiment the driver give a fixed steering wheel angle when driving the car with data acquisition system. The second experiment is introduced following; a skilled driver has driving the experimental vehicle to finish a series of parallel parking in a general space or in a tight spaces. This paper deals with the experimental data to get better fuzzy rules.In order to improve the fuzzy controller designed by subjective recognize, based on Adaptive-network-based Fuzzy Inference System (ANFIS), the experiment data has been analyzed to build fuzzy controller of the general space and tight space. According to the data, the starting area can be determined. the fuzzy logic comes from the analysis of the data from the actual driving, so it meets the basic requirements of actual parking. Then this paper also describes the advantages and disadvantages of fuzzy control and path planning.This paper builds parking experimental and reverse model in Matlab/simulink and put the control strategies based on path parking and fuzzy logic into the reversing system built in Matlab/simulink. From the result of simulation, when the car reverses from the starting area, the car under the control of the strategies could park into the space with velocity less then 5km/h. The simulation has confirmed enforceability effectiveness inclusive of the velocity fluctuation of the controller strategies. In terms of the controller strategy based on path-parking, when it make the 2-arcs-tangent curve as the reversing path, the minimum size of the parking space is (6×2.4);When tracking the 2-arcs-tangent trajectory, the steering wheel angle's value occurs a big jump, which make the control operations implemented more difficultly. When tracking the arctangent trajectory through turning steering wheel, the minimum size of the parking space is (6 . 5×2.5).The good or bad of the control strategy based on fuzzy logic largely depends on driving skills, proficiency and fatigue of driver. Flexibility and adaptability of the control strategy based on fuzzy logic is better than the path-parking's.At last, this paper summarizes main content of the full text and points out research direction and focus of this topic for future.This paper studies the algorithm of automatic parallel-parking steering control of the domestic A0 Class car from practical problems, major innovations are as follows:This paper estimates the size of parallel parking space based on the limit turning radius of the domestic A0 Class car.Based on reversing characteristics, this paper plans a 2-arcs-tangent curve as reversing trajectory. Analysis of the possibility collision, it determines the reversing staring area and solves the problem of velocity fluctuations during reversing, depending on the different speed to correct parking space.The continuous function of arctangent curve is used to fit the original 2-arcs-tangent curve, which solves the problem like steering wheel angle's change and rotating speed is too big and fast. Late in the reverse, the original steering wheel angle is improved by increased by small values time and time again to adjust heading angle of the body, which make the steering wheel angle's change rate moderate.This paper summarizes subjective experience of a skilled driver to refine fuzzy rules, which solves the parallel parking's reversing of a real vehicle.Based on ANFIS, this paper refines the fuzzy logic of parallel parking from the reversing experimental data,then acquires the steering controller of automatic parallel parking. |