Fast robot kinematics modeling by using a parallel simulator (PSIM)

High-speed computers are strongly needed not only for solving scientific and engineering problems, but also for numerous industrial applications. Such applications include computer-aided design, oil exploration, weather predication, space applications and safety of nuclear reactors. The rapid development in VLSI technology makes it possible to implement time consuming algorithms in real-time situations. Parallel processing approaches can now be used to reduce the processing-time for models of very high mathematical structure such as the kinematics molding of robot manipulator. This system is used to construct and evaluate the performance and cost effectiveness of several proposed methods to solve the Jacobian algorithm. Parallelism is introduced to the algorithms by using different task-allocations and dividing the whole job into sub tasks. Detailed analysis is performed and results are obtained for the case of six DOF (degree of freedom) robot arms (Stanford Arm). Execution times comparisons between Von Neumann (uni processor) and parallel processor architectures by using parallel simulator package (PSIM) are presented. The gained results are much in favour for the parallel techniques by at least fifty-percent improvements. Of course, further studies are needed to achieve the convenient and optimum number of processors has to be done