Variable-Frequency Imitation Learning for Variable-Speed Motion

TL;DR

This paper introduces Variable-Frequency Imitation Learning (VFIL), a new approach that trains models at variable sampling frequencies to better imitate variable-speed motions, improving accuracy and success rates.

Contribution

The paper presents VFIL, a novel imitation learning method that handles variable sampling frequencies and speeds, addressing limitations of traditional constant-frequency models.

Findings

Improved velocity-wise accuracy at interpolated and extrapolated frequencies

12.5% increase in overall success rate

Effective for variable-speed motion imitation

Abstract

Conventional methods of imitation learning for variable-speed motion have difficulty extrapolating speeds because they rely on learning models running at a constant sampling frequency. This study proposes variable-frequency imitation learning (VFIL), a novel method for imitation learning with learning models trained to run at variable sampling frequencies along with the desired speeds of motion. The experimental results showed that the proposed method improved the velocity-wise accuracy along both the interpolated and extrapolated frequency labels, in addition to a 12.5 % increase in the overall success rate.