Research On Trajectory Tracking Control Of An Unmanned Vehicle Based On Model Predictive Control

In recent years,with the rapid development and widespread application of driverless technology,the key technologies of driverless have also attracted a large number of domestic and foreign scholars to make the related researches.On the one hand,the research on driverless technology provides a theoretical foundation for the development of modern intelligent vehicles.On the other hand,the increasingly mature driverless technology can greatly improve vehicle ride comfort and driving safety,while providing technical support for the development and application of intelligent transportation system.Trajectory tracking is one of the key difficult technologies in the development of unmanned vehicles,and accurate trajectory tracking has been recognized as one of the core technologies for unmanned vehicle’s motion control in the automotive industry and academic fields.Compared with other control algorithms,the model prediction control(MPC)algorithm has the distinctive features of model prediction,rolling optimization and online correction.Therefore,this paper carries out the research on the trajectory tracking control of unmanned vehicles based on the MPC algorithm.The main contributions of this paper are summed up as follows:(1)A two-degree-of-freedom kinematic model of an unmanned vehicles is established based on the kinematics principle to provide control objects for two types of the desirable improved model prediction controllers.(2)A comprehensive linear time-varying model predictive controller(LTV-MPC)design method is proposed for the unmanned vehicle.First,the error model of the vehicle trajectory tracking system is constructed based on the established 2-DOF kinematics model of the unmanned vehicle.Then the linear parametric theory is used to discretize this 2-DOF model,and the design of the trajectory tracking controller is transformed into an optimization problem of linear quadratic programming within the framework of MPC.Finally,the effectiveness of the designed controller is verified in MATLAB/Simulink platform.(3)To further reduce the complexity of the controller design,a nonlinear model predictive control strategy(NMPC)is proposed.First of all,the real road is transformed to a nonlinear mathematical problem according to the 2-DOF kinematic model of the unmanned vehicle.Then,the nonlinear problem is solved by Euler method,and the future state is predicted through the nonlinear model,the current state quantity,as well as the control quantity sequence in the control time domain.Finally,the validation of the proposed NMPC controller is performed in MATLAB/Simulink platform.(4)An outdoor field experimental platform is built according to the experimental requirements of unmanned vehicles,which consists of a scaled real racing car,sensors,single-chip microcomputer,the developed controllers and the host computer.The designed controller is developed based on Arduino development board and MATLAB programming,and the effectiveness of the proposed LTV-MPC and NMPC controllers is verified in real vehicles under different preset path trajectories.