Model-agnostic Meta-learning for Adaptive Gait Phase and Terrain Geometry Estimation with Wearable Soft Sensors

2025-08-04

AI Summary

Fabric-based flexible sensors suffer from strong nonlinearity due to hysteresis, deformation, and wear-related misalignment, alongside challenges in cross-subject/terrain generalization and scarcity of calibration data. To address these, this paper proposes the first model-agnostic meta-learning (MAML) framework for joint estimation of gait phase and terrain geometry (e.g., slope angle). Leveraging a transferable deep neural network initialization, the method enables rapid adaptation to new users with only a few calibration samples. It achieves high-accuracy, simultaneous estimation across unseen subjects and complex terrains. Evaluated on data from nine subjects across five terrain types, the approach significantly outperforms baseline methods in gait phase prediction, locomotion mode classification, and slope angle estimation. It demonstrates superior cross-subject generalizability, faster fine-tuning convergence, and enhanced robustness against sensor nonlinearity and inter-subject variability.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Calibration & Uncertainty Quantification
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

This letter presents a model-agnostic meta-learning (MAML) based framework for simultaneous and accurate estimation of human gait phase and terrain geometry using a small set of fabric-based wearable soft sensors, with efficient adaptation to unseen subjects and strong generalization across different subjects and terrains. Compared to rigid alternatives such as inertial measurement units, fabric-based soft sensors improve comfort but introduce nonlinearities due to hysteresis, placement error, and fabric deformation. Moreover, inter-subject and inter-terrain variability, coupled with limited calibration data in real-world deployments, further complicate accurate estimation. To address these challenges, the proposed framework integrates MAML into a deep learning architecture to learn a generalizable model initialization that captures subject- and terrain-invariant structure. This initialization enables efficient adaptation (i.e., adaptation with only a small amount of calibration data and a few fine-tuning steps) to new users, while maintaining strong generalization (i.e., high estimation accuracy across subjects and terrains). Experiments on nine participants walking at various speeds over five terrain conditions demonstrate that the proposed framework outperforms baseline approaches in estimating gait phase, locomotion mode, and incline angle, with superior accuracy, adaptation efficiency, and generalization.

Problem

• Estimating gait phase and terrain geometry with wearable soft sensors
• Addressing nonlinearities from sensor hysteresis and deformation
• Adapting to new users and terrains with minimal calibration data

Innovation

• Model-agnostic meta-learning for adaptive estimation
• Fabric-based soft sensors for comfort and accuracy
• Efficient adaptation to new subjects and terrains