Trajectory Tracking Control System Of Unmanned Vehicle Based On Monocular Vision

The trajectory tracking control system of unmanned vehicle plays an important role in improving traffic safety,reducing environmental pollution caused by frequent starting and stopping of vehicles,and alleviating driver fatigue driving.With the in-depth study of the pilotless technology,the high efficiency and low cost trajectory tracking control system has become the key technology for the intelligent and practical unmanned vehicles.Compared to satellite navigation and inertial navigation,monocular vision navigation can be applied to open and uncovered conditions and has better economy when tracking trajectory in urban structured roads.The trajectory tracking control system based on monocular vision navigation is studied in the unmanned vehicle.In logic design,the architecture of LCM real-time communication system is built.According to the research content,the trajectory tracking control system is divided into four parts: lane detection module,trajectory tracking control module,underlying control module and vehicle execution module,and the problems to be solved by each module are defined and the communication protocol interface is determined.In the scheme realization,1)according to the lane detection module,the monocular camera is selected as the vehicle visual sensor.The lane detection and the lane center line track are used in the way of image preprocessing,lane line edge point set detection and three B spline curve fitting,thus the visual navigation process is realized.2)According to the trajectory tracking control module,the model predictive control algorithm based on vehicle model and the pure tracking control algorithm are studied through the construction of the vehicle kinematics and dynamics model,and the controller is designed to track the trajectory of the lane center line given by the lane detection module.Then,the controller controls the speed and the steering angle of the front wheel to realize the real-time control of the longitudinal and longitudinal attitude of the unmanned vehicle.3)According to the underlying control module,the control instruction given by the trajectory tracking control module is transferred to the various executor by the CAN network,and the unmanned vehicle from the perception layer to the decision control layer,and then to the complete trajectory tracking process of the executive layer is realized.4)For vehicle execution module,it consists of steering mechanism,motor,brake,transmission and other actuators.Finally,compared with the traditional trajectory tracking control system,a low-cost trajectory tracking control system is proposed by combining with the practical application of the project.By building a real vehicle simulation environment,the model constraint problem is verified in the model predictive control algorithm.By building a real vehicle experimental platform,the real vehicle verification of trajectory tracking control system based on pure tracking algorithm is realized.