Adaptive Identification of Legged Robotic Kinematic Structure

TL;DR

This paper presents an adaptive kinematic identification method for legged robots that accounts for deformation, improving model accuracy during high-impact actions like running and hopping.

Contribution

It introduces an algorithm that identifies kinematic structures from motion data, considering joint deformation, validated on simulation and real robot data.

Findings

Achieved 3.6% error in simulated kinematic identification

Confirmed joint deformation on ATRIAS robot during real tests

Predicted torques and forces using reconstructed joint models

Abstract

Model-based control usually relies on an accurate model, which is often obtained from CAD and actuator models. The more accurate the model the better the control performance. However, in bipedal robots that demonstrate high agility actions, such as running and hopping, the robot hardware will suffer from impacts with the environment and deform in vulnerable parts, which invalidates the predefined model. Thus, it is desired to have an adaptable kinematic structure that takes deformation into consideration. To account for this we propose an approach that models all of the robotic joints as 6-DOF joints and develop an algorithm that can identify the kinematic structure from motion capture data. We evaluate the algorithm's performance both in simulation - a three link pendulum, and on a bipedal robot - ATRIAS. In the simulated case the algorithm produces a result that has a 3.6% error compared to the ground truth, and on the real life bipedal robot the algorithm's result confirms our prior assumption where the joint deforms on out-of-plane degrees of freedom. In addition our algorithm is able to predict torques and forces using the reconstructed joint mode.