Statistical Modelling of Robot Manipulators

Abstract

A novel approach to the modelling of robot manipulators is presented. Usually, system descriptions are based on a set od differential equations which, due to their nature, lead to very precise results and strategies but, in general, require laborious computations. This motivates the need of alternative methods based on other mathematical concepts. The proposed statistical model gives clear guidelines towards the robot kinematic and dynamic optimization and the definition of rest and active manipulation regions. Furthermore, the results point out ideal properties for the actuators namely as "position and acceleration" devices instead of "velocity" machines. This observation is utmost importance as it gives a basis to new mechanical robot manipulator structures, with performances close to biomechanical systems.