Research On Adaptive Cruise System In Traffic Jam Environment Based On Visual Perception

In the context of the rapid development of computer vision technology and the strong demand and development of driver assistance systems,this article combines two advanced technologies and proposes research on stop&go control of vehicle adaptive cruise system in high-speed roads or urban closed roads with severe traffic jams.The ACC system architecture is based on layered play,includes the environment awareness layer,the kinematic decision layer and the actuator control layer.The environment awareness layer includes several key technologies such as vehicle recognition,distance detection,and taillight recognition.In this paper,a 104 ×104 scale detection feature layer is added to the original YOLOv3 network to further improve the accuracy of vehicle recognition.In this paper,the Gaussian optical model is used to derive the distance calculation method according to the pixel size of the license plate.The length and width of the extracted license plate are corrected and the camera is calibrated to improve the reliability and accuracy of the distance algorithm.The effective range of the method is 35 m,and the error is less than 2%.This paper uses area segmentation,HSV color space brightness and color threshold identification methods to identify various vehicle taillight conditions,which can meet the application requirements.The environment awareness algorithm used in this article has a single image processing time of 38.47 ms and can achieve basic 25 fps real-time measurement.The kinematics decision layer divides the driving mode of the vehicle into the following mode and the cruise mode,and the conditions for the mutual conversion of the two modes are designed.In the control strategy,the following mode adopts the constant time headway strategy considering the relative speed of the two vehicles and the driving state of the preceding vehicle,and uses the LQR control method based on the second-order longitudinal dynamic equation.The constant speed cruise mode adopts the PID control with the driving state of the preceding vehicle.The actuator control layer has designed the switching between drive mode and braking mode,as well as the control logic of starting,stopping and parking.The driving mode uses the vehicle longitudinal dynamic torque control method combined with the inverse engine model look-up table to convert the expected acceleration into the throttle opening value.And the braking mode uses the linear relationship between brake pressure and acceleration for calculation.This article uses a software simulation platform based on Carsim and Matlab/Simulink software to verify the simulation of the designed adaptive cruise start-stop control strategy,and compares the parameters in the simulation process to obtain the optimal value.The simulation results show that the adaptive cruise start-stop control system studied in this paper has sufficient safety and application value.