Research Of Visual Reversing Aid System Based On Backing Track Prediction Algorithm

Nowadays, with the popularity of the vehicle, more and more people have theirown cars. Reversing was the operation that were frequently used during driving. Butin real life, many people damage their car due to improper reversing, even whencircumstances have occurred resulting in more serious accidents. To solve thisproblem, parking assist system came into being, such as the common ultrasonicreversing radar, reversing video displays. However, these methods are only to someextent, reduce the difficulty of reversing the driver, guiding the driver and thereverse operation is not intelligent.In this paper, using two curve fitting idea designed a reverse trajectory predictionalgorithm based on visual parking assist system. Its real-time trajectory is predictedand shown under current steering angle, intelligently guided the driver operating thereversing and it will decrease vehicle damage or even the risk of traffic accidents.Paper introduces the reverse problem people encounter in daily life, that is thebackground parking assist system, a clear necessity and significance of its research,describes the working principle of common types of reversing the current auxiliaryequipment, and analysis and comparison their advantages and disadvantages, notedthat intelligently guide the reversing operation even autopilot technology will be thefuture direction of development.Then gives the derivation of the kinematic model of the car, calibrationprocessing by the camera based on it. By analyzing reversing path prediction methodsome of its shortcomings in the actual operation are pointed out. Highlighted themethod of least squares curve fitting based on theory, and a detailed explanation of theuse of the process to achieve its forecast reversing path.Subsequently, to test the predictive effect,the experimental test platform is described in detail. Mainly from the mechanical structure, as well as aspects of thehardware circuit control algorithm, the model describes the vehicle, the road controlmechanism, steering control, the master chip, reverse image acquisition camera,reversing video display terminals and wireless monitoring and control unit. From thechoice of the device, comparing the program and the achievement of the ultimatealgorithm are given a detailed description.Then, using the experimental test platform tests thereverse path predictionalgorithms. First written by Visual C++to get a reverse image from the hostcomputer and read the trajectory coordinate data. Using MatLab coordinate datasampling trajectory grouping fitting curves to obtain the parameters under the steeringangle stepper fitting. On this basis, Fitting the steering angle and the fitting functionof each group of parameters. Finally, use steering angle could predict thecorresponding trajectory. By comparing the predicted trajectory in real time with theactual trajectory is found that the trajectory prediction error is relatively small, and itis forecasting better.Finally, the results of the tests were analyzed and summarized, and the thinkingof possible improvements in the outlook of the system are decribed for futureresearch.