The Application Research Of Vision-based Intelligent Vehicle Control System And Algorithm

The main content of this article can be divided into platform structures, image acquisition, image data processing, control strategy, and restructuring other mechanical parts.Firstly, after being familiar with the features of each module according to their mechanical design and the characteristics be adjusted to the most optimal state, so that we can balance the focus of the whole car model and coordinate the proportion of the car.Secondly, Design the control circuit board, in-depth understanding each module of microprocessor MC9S12DG128, using the relevant image processing to extract feature and filter the data collected by the camera . reflect the characteristics of the image by figures and detect the edge, binary processing through the threshold, and identify the corresponding center coordinates. Through the appropriate algorithm to identify some special shapes, and then store the feature information in the microprocessor.Again,after the analysis of the image information acquisited by the camera, which provides a basis for the body, through the appropriate algorithm to design the control strategy of body motion, start and stop, acceleration and deceleration more reasonable.By comparing the Look-up table control strategy and fuzzy PID control strategy and through observation, the final choice is the fuzzy PID control strategy .Realised identification and optimization the straight track, curved track, operating results has been improved, especially in the corners the speed has greatly improved by ensuring a smooth body. The center of membership function selection is also very important , the center point directly impact on the stalls, and stalls contact with the motor speed, so diffrent centers of membership function have different Control effect. Self-learning ability of genetic algorithms allows the center value of membership function optimize to a right extent.Genetic algorithms are relatively strong in the global search ability, it do not need to determine the rules to control, coding simple and easy realise, so the system adjust the Gaussian membership function center by genetic algorithm.The design of intelligent vehicle can automatically start and stop according to data collected by the camera and do corresponding control, appropriate control algorithms ensure good forward-looking and robustness, which has achieved good results in actual tests.