Path Recognition And Control System Research Of Intelligent Vehicle Based On Machine Vision

The intelligent vehicle, involves various fields such as the environmentalperception, path planning, intelligent decision of driving behavior, vehicle navigationand positioning and vehicle automatic control. It is currently one of the most activefields in scientific research.The present paper, under the research background of the "Freescale Cup"Intelligent Vehicle Racing, analyzes the mechanical structure and the kinematicsmodel of the intelligent vehicle, improves the mechanical structure and designs thehardware circuit of the intelligent vehicle. And based on these, it researches the roadinformation extraction, processing algorithm and the control and decision algorithmof the intelligent vehicle.Firstly, this paper has improved the mechanical structure of the body of theintelligent vehicle. It mainly includes front wheel alignment, centre-of-gravityposition adjustment, ground clearance adjustment, steering engine installation andsensor installation. Secondly, this paper has analyzed and designed thehardware circuit of the intelligent vehicle. It consists of the following components:MCU core processing module, power module, motor driving module, steeringengine module, road information extraction module and speed detection module.Then this paper has described separately the structure and the basic workingprinciple of linear array CCD sensor and planar array CCD sensor. The two sensorsare used respectively for road information extraction and processing. And accordingto the processing results and external influencing factor, the feasibilityof linear array CCD can be determined. At last, this paper has proved the necessity ofcontrol strategy by establishing the kinematics model. When controlling the steeringengine and motor of the intelligent car by traditional PID Control Method, it isdifficult for the control parameter to self-tune with the change of circumstance, andthen the running status of the intelligent vehicle will be influenced. Concerning thisissue, the Single Neuron PID Control Method is presented. This method has strong self-tuning ability, little adjustable parameter, short period of regulation and goodstability. According to the Single Neuron PID Control Method, simulations ofseveral commonly used algorithms are tested. The compared results show thesuperiority of the control algorithms in this paper. And in the actualautonomous driving test of the intelligent vehicle, the feasibility of the SingleNeuron PID Control Method in the intelligent vehicle control has been proved.