A Study On Speed Profile Planning And Tracking Control For Autonomous Vehicles

As the vehicle-related industries continuously develop,the number of road vehicles has increased year by year,and road traffic accidents have become a pain point affecting people’s daily life.Fortunately,autonomous driving can greatly avoid the occurrence of traffic accidents or effectively reduce the severity of potential traffic accidents,and improve the driving safety and riding comfort of vehicles,so autonomous vehicles have been widely concerned by the academic and vehicular industry.Generally speaking,autonomous vehicles are composed of multiple subsystems such as environmental perception,sensor information fusion,decision planning and automatic control,including speed planning and tracking control.Therefore,it is of great academic value and engineering significance to carry out research on the speed profile planning and tracking control methods of autonomous vehicles.At present,the research difficulties are:how to ensure that the autonomous vehicle can successfully complete the speed profile planning and tracking control tasks when driving on a given road,and its algorithm can better meet the safety and comfort requirements.This paper studies the speed profile planning methods based on the optimization,which have both safety and comfort,and the speed tracking control method with great robustness.This paper’s main contents are:（1）This paper discusses three speed planning methods for autonomous vehicles:discrete speed profile sequence planning algorithm based on optimal quadratic programming,speed profile planning algorithm based on cubic natural spline curve and optimal quadratic programming,and multi-objective optimization speed profile planning algorithm based on quartic polynomial.The first two planning algorithms can transform the planning problem into a quadratic programming problem,while the third one can transform the planning problem into a"Max-Min"problem.By solving these problems,the final speed profile planning results can be obtained as the reference speed of the autonomous vehicle speed tracking control.In addition,the scene simulation tests verify that the speed profile planning results of the proposed three planning algorithms can meet the constraints,which are really effective and feasible.（2）Four speed tracking control methods for autonomous vehicles considering road slope observation are studied in this paper:the sliding mode speed tracking control algorithm based on the input-output feedback linearization（SMC）,the backstepping speed tracking control algorithm based on the input-output feedback linearization（Backstepping）,the backstepping sliding mode speed tracking control algorithm（BSMC）,the backstepping sliding mode speed tracking control algorithm based on linear matrix inequalities（BSMC-LMIs）.First,they all take the vehicle single-wheel longitudinal dynamics model as the nominal model.Considering that the slope observation can correct the nominal model,this paper proposes the road slope estimation method based on Extended Kalman Filter（EKF）,which introduces the nonlinear combined lagged slip tire models accurately calculate tire force.Secondly,this paper introduces the input-output feedback linearization method to linearize the nominal model.Moreover,BSMC-LMIs algorithm takes into account the extra tire lagged effect and lumped uncertainties（modeling error,parameter perturbation and external disturbance and noise,etc.）,and by solving the linear matrix inequality to derive a sufficient condition for the existence of the sliding mode surface,which ensures that the system trajectory in the sliding mode surface is t^-α asymptotically stable.Similar to BSMC algorithm,BSMC-LMIs introduces backstepping strategy can effectively reduce the chattering phenomenon inherent in sliding mode control.Finally,the robustness,feasibility and effectiveness of the proposed four controllers are verified by simulation tests under various typical working scenarios,which show satisfactory results when dealing with uncertainties on sloped roads.（3）Combined with speed profile planning and speed tracking control,a layered speed profile planning and tracking control system for autonomous vehicles is constructed in this paper,which consists of decision-making layer,planning layer and control layer.The decision-making layer conveys a decision signal to the planning layer,the planning layer decides whether to trigger the speed planning mechanism to obtain the desired speed according to the decision signal,and the upper controller of the control layer obtains the control signals（throttle or braking）in real time that make the autonomous vehicle drive safely and accurately according to the desired speed of the planning layer,while the lower controller manipulates the throttle or brake to control the vehicle to accurately track the desired speed.Finally,the simulation experiment verifies the effectiveness of the system,and show that the vehicle can successfully complete the driving tasks of speed planning and tracking control on a given path on the highway.