Dynamic optimization of N-joint robotic limb deployments

Abstract

We describe an approach using a stochastic optimization framework (SOF) for operating complex mobile systems with several degrees of freedom (DOFs), such as robotic limbs with N joints, in environments that can contain obstacles. As part of the SOF, we have employed an efficient simulated annealing algorithm normally used in computationally highly expensive optimization and search problems such as the traveling salesman problem and protein design. This algorithm is particularly suited to run onboard industrial robots, robots in telemedicine, remote spacecraft, planetary landers, and rovers, i.e., robotic platforms with limited computational capabilities. The robotic limb deployment optimization approach presented here offers an alternative to time-intensive robotic arm deployment path planning algorithms in general and in particular for robotic limb systems in which closed-form solutions do not exist. Application examples for a (N = 4)-DOF arm on a planetary exploration rover are presented.

Additional details