In recent years, mobile robot has been a research hotspot in robotics, developing towards the direction of autonomy and intelligence. It has represented the highest achievement of mechatronics. Compared to the traditional industrial robots, intelligent mobile robot with self-perception, decision-making and execution ability has getting extensive application in many fields especially in family, education, service and logistics industry. In this paper, omnidirectional mobile manipulator is studied, including autonomous navigation of the robot and autonomous grasping of the manipulator.The research was mainly carried out from the following aspects:Firstly, the youBot robot was selected as the research object, HOKUYO 2D laser range finder and Xtion 3D depth camera as sensor input. Then ROS software platform of robot was established. After kinematics modeling analysis of omnidirectional mobile platform, the position relation of the four-wheel motion and the moving platform in the ground coordinate system was derived for the preparation of the navigation of the mobile platform.Secondly, the SLAM technology for mobile robots was introduced. FastSLAM technology based on particle filter was used for autonomous navigation and A* algorithm for the path planning of the mobile platform. All above has been achieved in the software platform.Thirdly, modeling and analysis of the omnidirectional mobile manipulator was carried out. The forward kinematics of the manipulator was established by using D-H modeling method and its correctness was verified by MATLAB simulation. Meanwhile, the working space of the manipulator was obtained. Later writing C++inverse solution program to get all the inverse solution and verified it on the MATLAB. The optimal inverse solution was obtained according to the energy optimal principle. Then the omnidirectional mobile manipulator of 8 DOF kinematic models was set up. According to the workspace of the manipulator, three variables in the range of a mobile platform were established to provide target information for the navigation of mobile platform. What’s more, after solving the inverse kinematics of manipulator, the angle of each joint was obtained to prepare for the trajectory planning of manipulator.Finally, by connecting the laptop and the robot computer, the robot autonomous navigation experiment was carried out in the laboratory with barrier channel. The effect of navigation and obstacle avoidance was obvious, so the feasibility of SLAM algorithm and path planning algorithm was proved. After configuring the manipulator in MoveIt! simulation environment, when the manipulator enters the working range, the position and attitude of the target object can be identified by the Xtion depth camera calibration work. And a new solver for inverse solution program was wrote and configured to dealing with the problem that KDL solver failed to get inverse solutions of manipulator. Finally, the "pick and place" operation simulation and experiment have been achieved by using the mentioned new solver and path of robot trajectory planning based on quintic polynomial interpolation algorithm. The joint displacement, velocity and acceleration of the joint at the transition point were guaranteed to be continuous and smooth, which was an important preparation for the realization of the grasping of manipulator in a real environment. |