Specification and method for solving complex motion and calibration tasks

In this paper is presented an architecture permitting the specification and resolution of complex motion and calibration tasks. The description of such tasks leads to nonlinear functions optimization under equality and/or inequality constraints. The latter appears very commonly during robotics tasks execution. For example, while moving in a cluttered environment, it might be useful to keep some objects away from obstacles. The proposed method is based on a task description in terms of contact relationships to establish or maintain between parts of the modeled system. Some typical examples of tasks have been specified and performed to demonstrate the presented concepts. This architecture has been designed with the intention to integrating it into a remote manipulator teleprogramming system.