Development Of Regional Intelligent Electric Vehicle Experimental Platform Based On Visual Navigation

Intelligent electric vehicles not only reduce human driving fatigue,reduce the incidence of traffic accidents,but also reduce environmental pollution.Electric vehicles and their intelligence are the future development trends.It is of great significance to realize the autonomous navigation control of smart electric vehicles.Firstly,we build an Intelligent vehicles platform on the M300 electric vehicle.We build the upper computer interface based on the MFC,this interface can display the current traffic scene,deviation,the amount of control and vehicle front declination and other information.The lower machine controller adopts STM32F103ZET6 micro-controller to complete the development and testing of ADC,DAC,PWM,IO and other modules.The upper computer and the lower computer exchange data through the serial port.Using stepper motor as the power source for the automatic steering.The motor selection,installation,and driver program development were completed.Angle sensors were used to measure the angle and direction of the steering wheel.The steering transmission system was designed,engineered,simulated,and installed.Selected the XP70 HV servo as the power part of the brake mechanism to complete the driver development and debugging.Using single-chip DAC to achieve speed control.Secondly,extract the lane line.Transformed the grayscale image to the grayscale image and filtered the grayscale image by the median filtering algorithm.The edge of the image is detected.The logic and of the Canny edge detection map and the OTSU threshold map.Hough for logic and images,uses the k-means clustering algorithm to classify the Huff test results,and uses a least-squares method to fit the lane line equation.Thirdly,completed the path tracking controller design and simulation.A visual preview path tracking model is established,and the lateral deviation and azimuth deviation at the preview point are used as the running deviation of the vehicle.Established a two-degree-of-freedom dynamic model for vehicles.A self-adaptive fuzzy PID controller was built under Simulink,and vehicle model module,trajectory generation module,and preview misalignment calculation module were added to obtain trajectory tracking results,and the results were analyzed.Finally,completed the actual vehicle test.A lane line with a step distance of 0.5m was set up,and the vehicle speed was constant at 5km/h.Experiments were conducted and the experiment achieved good results.