| Trajectory planning and tracking control methods play an important role in intelligent driving technology and are the core links to realize assisted driving.Aiming at the intelligent driving scenario where global positioning cannot be obtained through the Global Navigation Satellite System(GNSS),but relative positioning information can only be obtained through local lane line,this thesis proposes a set of intelligent vehicle trajectory planning and tracking control methods based on local lane line to realize assisted driving functions in this scenario,such as deviation correction,lane keeping,static obstacle avoidance and stopping,and following driving.First of all,in view of the unsatisfactory planning effect of the traditional path planning method based on local lane line in the road with large curvature,a path planning method based on the vehicle curve coordinate system is designed:(1)The definition of the vehicle curve coordinate system is given with reference to the Frenet coordinate system,and multiple smooth paths are obtained through Lattice sampling in the vehicle curve coordinate system,and then the candidate path set in the vehicle coordinate system is obtained through coordinate transformation;(2)Build an evaluation function for the candidate paths,and sort the candidate path set according to the value of the evaluation function;(3)Physically check the candidate paths in turn to obtain the optimal path,and the speed-related safety redundancy is set for the collision detection,which increases the safety margin of the collision inspection during highspeed driving and improve safety in high-speed obstacle avoidance conditions.Secondly,in view of the limitations of existing speed planning methods,a trapezoidal velocity planning method under multi-constraint conditions is designed.Considering the impact of environmental constraints such as static obstacles,dynamic obstacles,and road curvature on speed planning in driving scenarios,the adaptive maximum planning speed is obtained according to the curvature and lateral offset of the planned path,and the trapezoidal speed planning based on the virtual intermediate speed is used to generate the speed curve,reducing Planning speed during cornering and obstacle avoidance and achieving stable following.At the same time,in order to solve the shortcoming of the acceleration sudden change of the trapezoidal velocity planning method,the gradient descent smoothing algorithm is used to post-process the velocity curve.Then,in order to verify the effectiveness and feasibility of the planning algorithm,a suitable tracking control algorithm is used to realize the tracking of the planned trajectory:(1)The path tracking is realized by the model predictive control algorithm based on the kinematic model;(2)Speed tracking is achieved using a hierarchical control algorithm with a PID controller plus an offline calibration table.Finally,the co-simulation platform of Pre Scan,Simulink and Car Sim and the experimental platform are built to verify the algorithm in this thesis.Under the unobstructed condition,the average lateral acceleration of simulation and experiment is reduced by 29.6%and 46.4%,respectively,and the maximum lateral offset is reduced by 36.7% and 28.6%,respectively,indicating that the smart car using the algorithm in this thesis has higher comfort and safety.By analyzing the simulation and experimental results under different working conditions,it is proved that the trajectory planning and tracking control algorithm proposed in this paper can realize assisted driving functions such as offset correction,lane keeping,static obstacle avoidance and stop as well as following the car based on local lane line in structured roads. |
