Design And Implementation Of Outdoor Logistics Robot Trajectory Planning System

Outdoor logistics robots have the advantages of lightweight,high flexibility,and full automation.It is a key component of an intelligent logistics system and have important practical value.Its industrial applications have received extensive attention and research.Trajectory planning is the core of outdoor logistics robot motion performance.The algorithm not only needs to ensure that the robot avoids obstacles flexibly,but also needs to take into account the high-speed transportation of the robot.Trajectory planning plays a connecting role in the algorithm structure of mobile robots.However,there are few studies on trajectory planning for outdoor logistics robots,most of which are only applicable to specific scenarios,and the planning results cannot meet practical needs.Therefore,based on the main problems faced by outdoor logistics robots in actual operation,this paper comprehensively utilizes the advantages of various planning algorithms,including search,sampling and optimization,and designs a trajectory planning module that can cover various scenarios.The main research contents of this paper are as follows:(1)Analyze the similarities and differences between outdoor logistics robot,indoor mobile robot and self-driving passenger car when facing planning problems.Based on the real data,the main tasks and scenes of the outdoor logistics robot are divided into crossing scene and parking scene.According to the scenes,the corresponding trajectory planning module is obtained.(2)In the scene of road crossing,a method based on sampling is proposed.Firstly build a trajectory sampling set,by sampling the location of the sampling end point along the road reference line and estimating the time from the starting point to the end point.Then design trajectory quality and risk cost functions to filter the trajectories in the sampling set.Finally,output the optimal trajectory.(3)In the parking scene,an improved Hybrid A*algorithm is proposed to search out the path to the destination.Then allocate the speed of the path based on Minimum jerk algorithm.Finally,the CFS algorithm is used to further optimize the output trajectory to obtain the final parking trajectory.(4)Write programs for testing.Most of the algorithms in this paper are implemented by C++ and run in actual robots.They are tested on the Longhua Foxconn and Shenzhen Virtual University Park platform to get the data of the robots.Experiments show that the planning system proposed in this paper can cover the most of scenarios,the output trajectory is smooth and executable,and has a high efficiency.In addition,the algorithm can meet the requirements of different computing platforms by adjust the parameter.