Research On The Fast Realization Method Of Trajectory Planning And Tracking Control For Unarmed Vehicle

Unmanned vehicles have gradually become the focus of the global "smart travel" wave,and will play a vital role in reducing the rate of traffic accidents and improving travel efficiency.The "planning + tracking" local motion framework is used as a key part of the unmanned vehicle system.The important part,local trajectory planning is to evaluate the current local surrounding environment information of the vehicle under the guidance of the global planning path,and comprehensively consider many factors such as safety,driving efficiency and ride comfort,so as to avoid obstacles and find a collision-free optimal Path;trajectory tracking control refers to real-time control of the optimal trajectory of the vehicle tracking plan,and satisfies the best evaluation indicators such as tracking accuracy or driving comfort.In addition,due to the complex and changeable driving environment of the vehicle,the local motion framework of the unmanned vehicle needs to complete the real-time optimization control task in a short time,which puts forward higher requirements for the calculation speed of the planning and tracking module.This paper studies the precise control and rapid realization of the "planning + tracking" double-layer controller.As follows:(1)In the vehicle local trajectory planning module,the researched local trajectory planning algorithm content consists of dynamic window search and local optimization.The screening consists of two parts,the dynamic window method is used to sample the dynamic trajectory clusters under the current and predicted windows,and the environmental potential field designed by the analogous artificial potential field method is used to screen out the optimal sub-trajectory,so as to find a collision-free optimal local trajectory;Finally,combined with the Car Sim real vehicle model,four simulation conditions of random obstacle avoidance,double-tracking winding piles,normal overtaking and failure overtaking are set,and the trajectory planning algorithm combining dynamic window method and artificial potential field method is verified by simulation..Analyzing the simulation results,it is found that the planning algorithm can efficiently and reasonably plan the feasible trajectory of the local environment for different working conditions.(2)In the vehicle nonlinear MPC trajectory tracking module,a nonlinear MPC trajectory tracking controller is built based on the theoretical knowledge of the model predictive control algorithm.The predictive model is a nonlinear vehicle dynamics model,and the objective function includes accuracy and stability.Based on the MATLAB optimization toolbox,the optimal control value of the optimization control problem is solved,that is,the front wheel steering deflection angle of the two-degree-of-freedom vehicle model;finally,the single-shift line,double-shift line and serpentine are set Line three simulation conditions,and four cruise speeds under each operating condition,observe the lateral tracking accuracy of the module,and create an incremental linear quadratic programming trajectory tracking controller as a control experiment.The results show that,compared with the quadprog controller of the control group,the constructed nonlinear MPC trajectory tracking controller not only improves the tracking accuracy by more than 10 times,but also ensures a more stable vehicle state.(3)In the fast solution algorithm for the vehicle’s nonlinear MPC trajectory tracking problem,based on the nonlinear MPC trajectory tracking controller,the MATLAB optimization solution toolbox is abandoned,the Pontryagin principle of extreme value is applied,and comorphic variables are introduced to construct The Hamiltonian function under the specific trajectory tracking problem,and according to the optimal control principle,the specific explicit optimal control rate and the dichotomy iteration rule of comorphic variables are obtained;then the general nonlinear trajectory tracking controller created in the previous chapter As a control experiment,design the same controller parameters and simulation conditions,and focus on the solution speed and control accuracy of the controller.The results show that the constructed trajectory fast-tracking controller can ensure that the vehicle tracks the desired trajectory more accurately,and at the same time,the solution time for the control quantity at each moment is reduced by nearly 370 times.(4)In the fast implementation algorithm of the "plan + tracking" local motion framework,the upper "trajectory planning" controller embeds the trajectory planning algorithm to plan a discretely distributed predicted trajectory sequence,and the lower controller embeds the fast trajectory tracking algorithm,Quickly calculate the front wheel deflection angle control amount,and transfer it to the vehicle state update model in the form of output;then optimize the prediction time domain and sampling frequency relationship of the upper and lower layers of the frame;finally,design dynamic normal overtaking simulation conditions and dynamics respectively Failed overtaking simulation conditions verify the validity of the framework.The results show that the constructed double-layer controller can control the vehicle to avoid obstacles and achieve the requirements of safe driving.At the same time,the solution time of the control quantity at each moment in the process is greatly reduced,which is conducive to real-time optimization control.