A NEW PROBABILISTIC METHOD FOR THE PERFORMANCE EVALUATION OF MANIPULATORS

A new probabilistic method is introduced which can be used to obtain a global map of the distribution of the velocity and force transmission characteristics of robot manipulators. The method yields the global d istributions of volumes of the velocity and force ellipsoids. It utili zes the most basic definition of the properties of the mapping functio ns between the joint coordinates and the Cartesian (task) coordinates. The method is extremely simple, numerical in nature, and readily impl emented on a computer. It requires very simple and limited analytical formulations. No inverse calculations are involved. The method can be utilized for optimal positioning of task spaces within workspaces of m anipulators. Velocity and force transmission (alone and in combination ) may be used as optimality criteria. Complex task space and manipulat or workspace geometries are readily handled. The tasks, velocity, and force transmission requirements may assume any arbitrary, even conditi onal, distributions within the task space. These distributions, especi ally conditional distributions, are extremely difficult, if not imposs ible, to handle using other methods. The application of the method to task planning and optimal manipulator synthesis is discussed. Several examples are presented.