Optimizing Force Signals from Human Demonstrations of In-Contact Motions

2025-07-21

AI Summary

In human demonstration, force signals during contact motion suffer from high noise levels and inaccurate first-contact point estimation, leading to erroneous intent recognition and distorted motion reproduction. To address these issues, this paper proposes a force signal optimization method integrating adaptive peak detection with multi-stage filtering: peak detection first corrects temporal misalignment of the initial contact event, followed by combined time–frequency domain filtering to suppress noise. A dedicated error metric is further introduced to quantitatively evaluate optimization performance. Experimental results demonstrate that, under this metric, the signal quality of single-action demonstrations improves by up to 20%. Consequently, the robot’s accuracy in inferring human operational intent and fidelity in reproducing demonstrated motions are significantly enhanced. This advancement improves both the reliability and usability of human–robot collaborative programming.

Technology Category

Intelligent Robots: Learning & Optimization for ROBHumans and AI: Planning and Decision Support for Human-Machine TeamsComputer Vision: Motion & Tracking

Application Category

Responsible Web: Machine-in-the-loop, human agency and autonomyEconomics, Online Markets and Human Computation: Data quality aspects of human-annotated datasetsUser Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

For non-robot-programming experts, kinesthetic guiding can be an intuitive input method, as robot programming of in-contact tasks is becoming more prominent. However, imprecise and noisy input signals from human demonstrations pose problems when reproducing motions directly or using the signal as input for machine learning methods. This paper explores optimizing force signals to correspond better to the human intention of the demonstrated signal. We compare different signal filtering methods and propose a peak detection method for dealing with first-contact deviations in the signal. The evaluation of these methods considers a specialized error criterion between the input and the human-intended signal. In addition, we analyze the critical parameters' influence on the filtering methods. The quality for an individual motion could be increased by up to SI{20}{percent} concerning the error criterion. The proposed contribution can improve the usability of robot programming and the interaction between humans and robots.

Problem

• Optimizing noisy force signals from human demonstrations
• Improving robot programming via better signal filtering
• Reducing errors in human-intended motion reproduction

Innovation

• Optimizing force signals from human demonstrations
• Comparing signal filtering methods for noise reduction
• Proposing peak detection for first-contact deviations