| China is a large fruit production country.With the continuous expansion of fruit planting scale,the demand for intelligent operation equipment that can replace human resources has become more urgent.Orchard mobile robot is a kind of mobile platform which can walk independently in the orchard.It can complete specific tasks in the orchard with different end actuators.This paper mainly aims at the complex and relatively fixed structure of orchard environment,and studies and realizes the key technologies of trunk recognition,autonomous navigation and obstacle avoidance of mobile robot in orchard without establishing global electronic map in advance.The specific contents of this thesis are listed as follows:(1)The functional requirements of orchard mobile robot are analyzed,and the research technology route is determined.According to the function of each module,the hardware parts such as sensors are selected,and the experimental platform is built.The key functional components of the hardware system are analyzed,and the software development environment used in this research is introduced.(2)Aiming at the function of tree trunk detection in orchard with weeds and slopes,a tree trunk detection algorithm based on Adaptive Density Clustering is proposed.For the problem that DBSCAN clustering algorithm is sensitive to global parameters,combined with the scanning characteristics of single line lidar,clustering threshold is set adaptively.Then,according to the data characteristics of the ground and the weed branches and leaves,we can eliminate them and complete the detection of the trunk.The experimental results show that the algorithm is suitable and robust for the detection of tree trunks with a small amount of weeds and slopes.(3)In view of the autonomous navigation and walking function of orchard mobile robot,the line of detected fruit trees is extracted by Hough transform,and the center line of tree row is used as the navigation path of mobile robot.And the PID theory is used to control the mobile robot.The simulation results show that when lidar scans the trunk and the crown,the line extraction algorithm has better performance.The results of real vehicle experiments show that the lateral deviation of the two cases is no more than 21cm,which can meet the basic requirements of driving in a straight line in the orchard.(4)In view of the obstacle avoidance function of orchard mobile robot,the traditional artificial potential field method is used as the basis of obstacle avoidance algorithm,and the target influence factor is added to the repulsion potential field function,and then the principle of parameter selection of the improved artificial potential field method is analyzed.Finally,the simulation and real vehicle experiments are carried out.The results show that the improved artificial potential field method can effectively avoid the local minimum problem after selecting reasonable parameters,and can achieve the requirements of obstacle avoidance function.This paper can provide some theoretical support for the research of key technologies such as navigation and obstacle avoidance of orchard mobile robot. |
PDF Full Text Request