Identification of Motor Parameters on Coupled Joints
This work addresses a specific challenge in robotics control for engineers and researchers, offering an incremental improvement in parameter identification for coupled joints with static friction.
The paper tackles the problem of accurately identifying motor torque and friction parameters in coupled joints, where traditional single-joint methods fail due to entanglement, and proposes a direct open-loop method that addresses each motor separately and accounts for static friction, resulting in improved accuracy compared to classical closed-loop methods.
The estimation of the motor torque and friction parameters are crucial for implementing an efficient low level joint torque control. In a set of coupled joints, the actuators torques are mapped to the output joint torques through a coupling matrix, such that the motor torque and friction parameters appear entangled from the point of view of the joints. As a result, their identification is problematic when using the same methodology as for single joints. This paper proposes an identification method with an improved accuracy with respect to classical closed loop methods on coupled joints. The method stands out through the following key points: it is a direct open loop identification; it addresses separately each motor in the coupling; it accounts for the static friction in the actuation elements. The identified parameters should significantly improve the contribution of the feed-forward terms in the low level control of coupled joints with static friction.