Deep Learning with Predictive Control for Human Motion Tracking

TL;DR

This paper introduces a novel approach combining deep learning and model predictive control to enhance human motion tracking accuracy in robotics, demonstrating significant performance improvements in handwriting tasks.

Contribution

It presents a new framework integrating MPC with deep learning and online DyBM learning for improved human motion tracking in robots.

Findings

Significant improvement in tracking accuracy.

Effective switching between learned and conservative predictions.

Successful application to handwriting motion tracking.

Abstract

We propose to combine model predictive control with deep learning for the task of accurate human motion tracking with a robot. We design the MPC to allow switching between the learned and a conservative prediction. We also explored online learning with a DyBM model. We applied this method to human handwriting motion tracking with a UR-5 robot. The results show that the framework significantly improves tracking performance.