Local Path Planning Method For In-park Intelligent Logistics Vehicles

At present,in-park intelligent logistics vehicle transportation is one of the hot scenes in the current autonomous driving theory research and technology application.Local path planning is the core technology of autonomous driving.The local path planning of intelligent logistics vehicles in the park mainly includes two scenarios: forward driving and parking.In view of the logistics and transportation scenarios in the park,the study of safe,stable and efficient local path planning methods is of great significance to improve the autonomous driving performance of logistics vehicles and the intelligent level of logistics transportation in the park.Therefore,this thesis develops forward path planning methods and parking trajectories of logistics vehicles planning method research.In order to plan a safe and stable local path in forward path planning,a vehicle threat assessment model that comprehensively considers lane lines and obstacles is established.Furthermore,based on the threat assessment model,a cost function that comprehensively considers the safety and stability of the trajectory is designed to find the optimal trajectory.Firstly,this thesis analyzes the risk factors of driving environment,and constructs a threat assessment model based on potential field with lane lines and obstacles as objects,which is used to describe the impact of lanes and obstacles on driving safety.Secondly,the fifth-order polynomial curve in the Frenet coordinate system is used for horizontal and vertical trajectory planning.State sampling strategies and trajectory evaluation cost functions under different working conditions are designed to generate horizontal and vertical candidate trajectories and evaluate trajectory quality.The horizontal and vertical candidate trajectories are synthesized and comprehensively evaluated according to the cost,and then the optimal trajectory is obtained.According to the characteristics of the transportation scene in the park,a planning response area is constructed to reduce unnecessary planning.Finally,through simulations of several typical scenarios,the effectiveness of the forward path planning in this thesis is verified.In the vertical parking scenario,in order to plan the optimal trajectory of the parking time under complex parking constraints,this thesis uses a numerical optimization method to transform the parking planning problem of the optimal control form into a discrete case of nonlinear planning optimization problem.In the process of solving nonlinear programming optimization problems,a partitioned and inverse solution method is proposed to reduce the difficulty of solving and the calculation time.First,this thesis establishes vehicle kinematics equations and analyzes the external constraints,internal constraints and boundary value constraints in the vertical parking process.Then,the time-optimal vertical parking objective function is established,and the parking problem is discretized by the orthogonal configuration finite element method.The objective function is solved by the partitioned and inverse solution method to obtain the optimal trajectory of vertical parking.Finally,through simulation experiments of different parking starting states,the effectiveness of the vertical parking trajectory method in this thesis is verified.According to forward path planning and vertical parking trajectory planning methods,this thesis designs and develops a local path planning software system,and verifies the effectiveness of the research method and software system in this thesis through actual vehicle experiments.