Research On The Following Control Simulation Of Driverless Vehicle

The development of driverless cars is of great significance to road traffic safety and can effectively reduce accidents caused by human factors.Following control is one of the important operating conditions for driverless driving,which can make vehicles run in an orderly manner and maintain intervals,thereby improving traffic utilization,ensuring smooth roads and avoiding rear end collisions.Therefore,in this paper,through research on front vehicle recognition,sensor information fusion,and follow-up control algorithms,we can ensure the safe driving of unmanned vehicles when following the control conditions.Use a monocular camera to collect the vehicle information in front,and use YOLOv3 neural network detection algorithm to perform vehicle detection and positioning on the image collected by the camera;use local binary feature extraction and support vector machine classification to identify the vehicle information on the rear Vehicle rear recognition;at the same time,the speed and distance of the preceding vehicle should be measured by the laser radar sensor.Fusion of image information and point cloud information from two aspects of space and time,complete the spatial fusion by transforming the camera and lidar coordinates,image coordinates,and world coordinates;complete the time fusion by adjusting the sampling interval of the two sensors;The R-Tree spatial index algorithm is used to quickly match the distance information of the point cloud in the recognition frame.Establish a safety distance model based on the fixed head time algorithm and calculate the safety distance in real time;complete the throttle and brake switching through the established throttle brake switching model;use the safety distance as the desired distance and design a follow-up control strategy;design a fuzzy PID controller And simulation,complete the acceleration and deceleration control of the vehicle.The data obtained from front car identification and information fusion was imported into Carsim software and co-simulated with the established Simulink control model.Five operating conditions were set: front low speed and low speed follow,front low speed and high speed follow,front high speed and high speed follow,and front low speed change.Follow low speed,follow high speed and follow high speed.Simulations under these five operating conditions yield five indicators: vehicle speed,safety clearance,acceleration,throttle opening,and brake pressure.The results show that the designed driverless following closed-loop control system has accurate identification and fusion,and reasonable control spacing.,Can complete the follow-up control of driverless cars.