Switching algorithm for fast robotic tracking under joint speed constraints

In the great majority of commercial robots the tracking speed along the task-space path usually has to be computed a priori by the robot operator to avoid hitting some input constraints. In this paper, an approach based on sliding mode conditioning of a path parametrization is proposed to achieve the greatest tracking speed which is compatible with the robot input constraints (e.g. joint speeds). The proposal avoids on-line computations required by conventional analytical methodologies, and it can be easily added to pre-existing path tracking schemes, since its switching law is confined to the low-power side of the system. A necessary and sufficient condition for sliding mode establishment is derived, which assures for a given path that the robot fully exploits the actuator power once sliding regime is reached. Simulation results on a 2R manipulator are presented in order to illustrate the effectiveness of the approach.