| A systematic method for the efficient and comprehensive simulation of the dynamics of robot manipulators has been created. We are able to determine the motion of any open-chain, serial-link robot with arbitrary geometry. Our simulator includes a number of phenomena which are known to influence the dynamic response of the manipulator; such as joint compliance (including gear train deformations), geometric errors, power amplifier nonlinearities, and actuator saturation. We model the kinematic structure of the manipulator by the ""modified Paul-Rosa representation"" (MPR), which utilizes an internal algorithm to ""extract"" the kinematic parameters from the convenient ""Paul-Rosa representation"". The MPR method compares favorably with the commonly used ""Denavit-Hartenberg notation"" in terms of computational efficiency and user-friendliness. Many by-products of the efficient kinematic analysis are directly applied towards the formulation of the dynamics equations of motions, via the d'Alembert-Lagrange virtual work approach.; Based upon the theories developed, a package of computer programs has been created. The major program: ""DYROB"" (DYnamics of ROBot manipulators), is capable of generating the time history of each joint motion as well as the displacement, velocity, and acceleration for each point of interest. The motion of the manipulator is graphically animated by the program ""ANIROB"" (ANImation of ROBot dynamics). Another program called ""DYROBIN""(DYROB INput program) is aimed at assisting the user in the preparation of input data files.; A control policy based on the position and velocity feedback of each individual joint has been incorporated into the program. To optimize the performance of the system, we implemented an algorithm which is capable of adjusting the controller gains according to the variation in joint positions and gravity loading conditions. The resulting accuracy in position tracking is quite satisfactory. The user of the program can easily adopt any other control scheme applicable to the specific task requirement.; A number of examples have been worked out to demonstrate the capability and performance of the simulation. In each case, a different manipulator is chosen for a specific task; parametric analysis of the errors is performed, and results are presented. |
