Kinematic Modeling And Trajectory Planning For A Variable Geometry Truss Manipulator

Variable Geometry Truss(VGT) manipulator is a kind of hyper-redundant manipulator. Compared with the traditional manipulators, this kind of manipulator has many advantages such as stiffness, modularity, flexibility which make VGT manipulator can be used in space-based remote operations, underwater detection without people, industrial environment with obstacle and soon. Based on the concept of VGT manipulator, this paper introduces a kind of distinctive manipulator which is composed by two symmetrical octahedral VGT. The manipulator introduces modular thinking and make kinematic problem for the whole manipulator simplify to kinematic problem for single module. By considering the symmetry in the modeling process for the manipulator, we can deduce the solution for the kinematic problem by using the geometry relationship of the manipulator.To calculate the control variables of the multiple-module manipulator, we design two different point to point trajectory planning algorithms from different starting points based on the kinematic of the manipulator. The first algorithm is based on the requirement for the project. By considering the basic geometry relationship of the manipulator model and by further simplifying the manipulator model, the trajectory planning problem is converted to an unconstrained optimization problem, so by using BFGS algorithm, the algorithm can be solved. The second algorithm is concentrated on the theory research and introduces the obstacle-avoidance requirement to the problem. So we can convert the problem to a quadratic convex optimization problem and if some parameters are appropriately chosen, the proposed algorithm can be proved to be of global convergence and to have superlinear convergence rate. To verify the effectiveness of the proposed algorithm, we also give the simulation experiment in the laboratory and the validation experiments on the physical platform. The results show that the two algorithms can meet the needs of robot control to some degree.