| In our country,fruit production accounts for 55% of the world’s total,which shows that China has become a veritable fruit producing country.However,orchard management in China is still dominated by manual operations,and the intelligent level of orchard machinery is not developed enough.With the development of intelligent robotics,it can effectively improve the intelligence of orchard operation machinery by introducing robotic technology into orchard.Autonomous navigation is the premise of the application of intelligent robots in orchards,which is the basic function to ensure the robot to carry out multi-functional operations without human control.According to the requirements of autonomous positioning and navigation for mobile robots in orchard environment,some key technologies are studied in this thesis,including autonomous positioning,linear trajectory tracking and autonomous obstacle avoidance.The main work is completed as follows:(1)The development status of autonomous positioning and navigation technology for orchard mobile robot is reviewed.And the functional requirements of orchard robot positioning and navigation are analyzed.Considering functional requirements and application scenarios,the system framework of the software & hardware is designed.The hardware framework consists of the system integration and the selection of the key components,and the software framework is based on the host computer interface.Additionally,the coding development environment is briefly introduced.(2)Aiming at the problem of losing GNSS differential signals in orchard environment,an algorithm of Li DAR odometry based on GICP matching is studied to assist the robot positioning and navigation.In the algorithm,we firstly employ a lidar to obtain the three-dimensional dense point cloud in orchard environment.Secondly,the pose transformation of the robot is analyzed by GICP matching,and the local submap is generated by adaptive key frame system for rapid global pose optimization.Finally,the pose of the robot in orchard environment can be obtained based on the pose transformation matrix.(3)The obstacle avoidance strategy combined with the GNSS and Lidar odometer information is researched.This strategy consists of three steps.(ⅰ)Describing the positioning information of the GNSS and laser odometer in a unified coordinate system by the Gaussian projection and coordinate transformation.In this way,the unified description of robot positioning,navigation path and obstacle information in this coordinate can be obtained.(ⅱ)Establishing the robot kinematics model,and implementing the linear trajectory tracking based on PID control theory.(ⅲ)the local obstacle avoidance environment of the robot is described by grid map,and the obstacle avoidance function of the robot is realized by combining the obstacle avoidance algorithm based on artificial potential field method.(4)The autonomous positioning and navigation system of orchard mobile robot studied in this paper is tested.We conduct the field experiments in a real orchard environment to verify the effectiveness of our algorithms including the robot’s perception and positioning,liner trajectory tracking and obstacle avoidance functions.Experimental results show that the maximum lateral deviation of the robot during the navigation is under 8cm,and the average lateral deviation is under 3cm.As a result,we believe that the autonomous positioning and navigation system of orchard mobile robot studied in this paper has enough reliability and applicability,and basically meets the navigation requirements of orchard mobile robot. |
