Research And Implementation Of Motion Planning Of Humanoid Robot Based On ROS

With the growing problem of aging society and the rapid decline in the proportion of the labor force,the trend of robots replacing human resources to provide services is becoming increasingly evident in the social service industry where the demand for staff is huge.In the field of service robots,humanoid robots are preferred by the majority of families for their human-like movement patterns.They have huge commercial application value and have attracted experts and scholars worldwide to study them.In this thesis,the humanoid robot's motion planning is taken as the entry point,the robot operating system(ROS)is used as the algorithm design framework,and the humanoid robot NAO is used as the carrier.The path planning of the NAO robot from the starting point to the target point in the non-reconstruction environment and the movement planning of the NAO robot arm are completed respectively,and a set of movements are realized.The main works are summarized as follows:Firstly,the background,significance and development status of research in the field of service robots are analyzed in detail.The background of humanoid robot,robot operating system ROS,motion planning and control and the research status at home and abroad are described in detail,and the importance of the research of this subject is proposed.Secondly,the NAO robot motion platform was built using the 3D visualization tool Rviz,and motion planning tool MoveIt! in ROS.The basic composition of ROS system,the operating mechanism and principle of ROS,and the joint parameters of NAO robot was described.Under the ROS system,the NAO robot model file was used to realize the 3D visual display of the NAO robot model with the help of the visualization tool Rviz.Finally,the motion group node is compiled for the robot motion control,and the NAO robot motion platform under the ROS system was completed.Next,the RRT algorithm was employed to implement path planning on existing grid maps and NAO robots was used to follow the planned path.The constraint types of the NAO robot on the two-dimensional plane are analyzed and the parameters of the fast and random tree algorithm are determined.Analyze the requirements of path planning,analyze the characteristics of laser radar and depth sensors,and the characteristics of geometric feature maps,topological maps,and grid maps,and establish the use of laser radar to create a two-dimensional grid map.We deeply studied the RRT algorithm,established a map model to implement a RRT algorithm,and changed obstacles to view different planning paths.In the two-dimensional grid map,the path planning is implemented using the RRT algorithm,and the NAO robot is used to move along the planned path.Finally,the MoveIt! tool in ROS was used to realize the motion planning of NAO robot arm under different scenarios.The basic theory of arm movement was studied in depth,including the representation of the rigid body posture and the kinematics of the robot.MoveIt! was used to specify the target state of the left arm of the NAO robot on the menu interface,and use the OMPL motion planning library to realize the movement of the left arm from the initial state to the target state,and apply this planning path to the real robot.Through the MoveGroup interface,programming NAO robot arm movement planning.Customize the planning scenario,add collision objects in the environment,and implement motion planning for obstacle avoidance.