Research On Path Planning Of Redundant Manipulator Based On ROS

In recent years,robot technology has developed rapidly,and it has been widely used in various industries and different fields of society.Especially in some special and dangerous environments that are difficult for humans to operate,robots have played an important role.Traditional robots have long operation time,low efficiency,low accuracy and lack of intelligence.Nowadays,with the continuous development of intelligent technology,the robot technology in special environment is also facing an update,which is necessary to study the precision and intelligent control of the robot arm.With high flexibility,fault tolerance and reliability,the redundant manipulator can avoid obstacles,avoid singularity,avoid joint limit,and optimize joint torque by using its self-motion characteristics.However,there are few studies on redundant manipulators and available domestic and foreign references,which greatly increases the difficulty of the research.At the same time,there are often several inverse kinematic solutions of redundant manipulators.How to choose from several joint combinations is also a difficult point to study.In this paper,the open source operating system ROS(Robot Operation System)is used as the simulation platform for the problem that the traditional robot simulation platform is oriented to specific objects,the model structure is complex and the function is single,etc.The following research work was carried out for the path planning of redundant robot arms:First,the kinematics of the redundant artificial arm was analyzed,the structure of the robot arm was designed and simplified in essence,and the corresponding kinematics model was established by d-h method.Robotics toolbox in MATLAB software was used to analyze the kinematics of the robot,and the angular velocity of the joint initial point and target point was used as the constraint condition.The interpolation method was used to realize the joint space trajectory planning.The simulation verifies the effectiveness of the planning.Second,the basic principles of the ROS was introduced,and a robot simulation platform was built based on the ROS system.Use the URDF file to describe the three-dimensional model of the manipulator,and the kinematics configuration of the robot arm were carried out through Moveit!,the configuration and startup files needed for the robot arm planning were generated.Using C ++ and other related APIs provided by Moveit!,combined with RRT path planning algorithm,an obstacle avoidance planning experiment was carried out on the robotic arm on the platform,which solved the problem of selecting several sets of joint angles of the redundant robotic arm and the problem of poor obstacle avoidance and flexibility of non-redundant mechanical arms.At the same time,in view of the disadvantages that the grasping posture and the placing posture need to be set in advance in the traditional planning,the adaptive grasping of the robot arm is proposed to improve its flexibility.Finally,the motion planning method was applied to the robotic vision grabbing experiment.Firstly,a monocular vision mechanical arm grab system was constructed,and the internal and external parameters of the camera were calibrated.Then,cv＿bridge is used to convert the image message in ROS into the image format available to Open CV.Then,RGB channel separation method is used to distinguish the positions of different colors in the interface by using the gray value in each channel,so as to realize the recognition of the target object in Gazebo environment and the grab experiment of mechanical arm combined with vision.The results show that this method is simple and effective,and the control method designed based on ROS is highly portable.At the same time,it can solve the problem of heavy workload and poor portability in system design when the robot combines with machine vision to grasp a specific object.More to meet the future needs of the industry for the complexity of production and the flexibility it requires.