Abstract
Increasing energy efficiency in automation can lead to reduced operating costs and enhanced production sustainability. In this paper, we present a novel approach for energy efficiency in redundant robotic systems. The proposed scheme aims at determining the weighting matrix used to compute the minimum-energy solution of the inverse kinematics through the weighted Jacobian pseudo-inverse. Two different solutions for computing the energy-efficient weights are proposed, one independent and the other dependent on the desired robot end-effector trajectory. The proposed approach also accounts for robot dynamics uncertainties. The performance of the approach is validated on a robotic system with 8 degrees of freedom, composed of a manipulator mounted on a linear axis. The results of extensive numerical simulations and bespoke experimental tests demonstrate the effectiveness of the proposed approach in reducing the mechanical energy consumption with respect to a state-of-the-art approach. A reduction of the energy expenditure up to 97.5% is found in the numerical tests, and up to 97.2% in the experiments, while guaranteeing robustness to imperfect knowledge of the robot dynamics parameters.