Kinematics Analysis and Trajectory Planning of Mobile Manipulators by Yaolun ZHANG Thesis Supervisor Professor Yangmin LI Master of Science in Electromechanical Engineering This thesis firstly presents the proposed algorithm in general, then it is mainly divided into three parts: the first part is on the topic of the kinematics analysis for manipulators, and the second one is about the kinematics analysis for locomotors, finally, the orientation planning is described for the trajectory planning in the Cartesian space. The inverse kinematics is of fundamental importance for trajectory planning, motion control, and workspace analysis, and even for the dynamic control. Currently the inverse kinematics is classified for iterative numerical solution and closed-form solution. The former exists some obvious shortcomings, so the latter is preferred for industrial application. Although there exist many different methods for serial manipulators, they often cannot directly specify which one is the optimal solution among the multiple solutions, and they often need other algorithms to select and match from multiple solutions, which has a great influence on the computational efficiency. Additionally, as for orientation planning, there exist some methods, such as interpolation of quaternion, Frenet – Serret formulas, Axis – angle representation, Euler angles. The interpolation of quaternion is popular among them. However, all of them may meet the requirements of six-DOF manipulators, and they still encounters singularity for kinematically deficient manipulators, like KUKA youBot manipulator. To settle the above-mentioned problems, a novel inverse kinematics method is proposed with the reservation method, meanwhile, the dynamic reference orientation and static reference orientation are also proposed to plan the orientation. The general description of the trajectory planning in the Cartesian space is as follows: firstly, the coordinate frames are attached to links, then the forward kinematics can be obtained through the analytical methods of kinematics. According to the vector geometric properties and the desired configuration during trajectory planning, the joint angle of the controlling joint and middle joint can be calculated. With the calculated joint angles and the related formulas from the analytical methods of kinematics, other joint angles can be settled. What’s more, if the fine-tune joint and the self-optimizing (so) joint exist, the fine-tune method and the self-optimizing method are proposed to settle the related problems. When obstacles exist, the reachable pose can be predicted through vector geometric properties in advance. Additionally, the continuous time-varying orientation can be planned that is coupled with position. The above-mentioned algorithm can avoid extraneous solutions, can avoid selecting and matching the optimal solution, and can plan the non-singular path efficiently, what’s more, compared with the one planned in joint space, it could modify the complicated shape of end-effector’s path, etc.. Finally, all the algorithms above are verified through simulation, and the related experiments are carried out based on KUKA youBot.