Research On Active Obstacle Avoidance System Based On Braking And Lane Changing Control

With the increasing number of cars,the number of accidental and dangerous accidents involving vehicles has risen significantly,and the technology to ensure driving safety has received a lot of research and discussion.The active obstacle avoidance system help the driver brake the vehicle to avoid obstacles by detecting the collision risk of the vehicle,and the system can effectively reduce the incidence of vehicle collision accidents.The current active obstacle avoidance system mainly adopts Obstacle avoidance strategy of longitudinal braking based on a single safety distance model.The system cannot adapt to obstacle avoidance scenarios with higher speeds.In addition,current active obstacle avoidance systems lack a distinction between vehicles and vulnerable users on the road,and unable to adjust the best obstacle avoidance strategy for different obstacle targets.In view of the above problems,the thesis mainly studies the vehicle risk estimation method and obstacle avoidance strategy.Firstly,the overall strategy and risk estimation model of active obstacle avoidance are studied.By analyzing the functional requirements of the obstacle avoidance system and considering the best obstacle avoidance method in different driving environments,the overall strategy of the vehicle’s active obstacle avoidance system is established,and the system adopts the steering obstacle avoidance strategy to optimize the longitudinal obstacle avoidance.Dangerous target detection is the first step to achieve obstacle avoidance,and in order to obtain the accurate categories of road users in complex scenes,an obstacle detection model is built based on YOLOv3 algorithm after comparing and studying different deep learning target detection algorithms.The different judgment methods of main dangerous target for different obstacle categories are designed.In order to improve reliability of risk estimation model,a safety distance model based on sigmoid function weighting is established by studying the characteristics of the current safety distance model for obstacle vehicles.Secondly,vehicle longitudinal obstacle avoidance control is studied.In order to accurately control the longitudinal state of the vehicle,the vehicle’s longitudinal dynamics model and reverse braking model is built by analyzing the vehicle dynamics,and uses the interpolation method to build an inverse engine model based on the actual engine map data,which is used to calculate the throttle opening corresponding to the system acceleration demand.In addition,considering that the system operation conforms to the real driver as much as possible,a hierarchical braking strategy of longitudinal dangerous vehicle targets and pedestrian targets is established based on real driver braking data and different vehicle dangerous states.Finally,a PID acceleration controller is built based on the vehicle dynamics model to speed up the vehicle’s response to the system braking demand.Then,vehicle steering obstacle avoidance control is studied.A reasonably simplified vehicle linear two-degree-of-freedom model is built by means of analyzing the vehicle’s lateral control requirements.In addition,considering the impact of vehicle tires on vehicle state control,the simulation model of tire is built based on the magic empirical formula.In order to formulate a reasonable steering obstacle avoidance path,by studying the steering trajectory planning algorithm,a steering obstacle avoidance path model is established based on fifth-degree polynomial fitting.Finally,a steering obstacle avoidance tracking controller based on model predictive control is established based on the vehicle linear two-degree-of-freedom model to reduce the tracking error of the planned path.Finally,referring to the C-NCAP management rules and European Euro-NCAP test standards,simulation scenarios of vehicle active obstacle avoidance are established based on Prescan,and Pre Scan and Simulink are used to jointly simulate the established active obstacle avoidance system.The simulation results show that the active obstacle avoidance system designed in the thesis has good obstacle avoidance performance,and the system can control the vehicle to effectively avoid collisions when the vehicle speed is lower and higher.