Bridging Structural Dynamics and Biomechanics: Human Motion Estimation through Footstep-Induced Floor Vibrations

TL;DR

This study tackles non-intrusive estimation of lower-limb joint motion using footstep-induced floor vibrations by marrying gait biomechanics with structural floor dynamics in a physics-informed graphical model. The proposed PIG system encodes joint, time, vibration, and body information with edges representing physiological and dynamic relationships, and enforces physical constraints via structure-property learners and biomechanics-inspired transforms. In a real-world 20-subject experiment across four gait types, the method achieves a mean absolute error of $3.7^{\circ}$ for 12 joint flexion angles, about a $38\%$ improvement over a tuned LSTM baseline and competitive with camera- and wearable-based approaches. The work demonstrates a non-intrusive, device-free pathway for gait health monitoring in daily living spaces, with implications for early disorder detection and fall prevention.

Abstract

Quantitative estimation of human joint motion in daily living spaces is essential for early detection and rehabilitation tracking of neuromusculoskeletal disorders (e.g., Parkinson's) and mitigating trip and fall risks for older adults. Existing approaches involve monitoring devices such as cameras, wearables, and pressure mats, but have operational constraints such as direct line-of-sight, carrying devices, and dense deployment. To overcome these limitations, we leverage gait-induced floor vibration to estimate lower-limb joint motion (e.g., ankle, knee, and hip flexion angles), allowing non-intrusive and contactless gait health monitoring in people's living spaces. To overcome the high uncertainty in lower-limb movement given the limited information provided by the gait-induced floor vibrations, we formulate a physics-informed graph to integrate domain knowledge of gait biomechanics and structural dynamics into the model. Specifically, different types of nodes represent heterogeneous information from joint motions and floor vibrations; Their connecting edges represent the physiological relationships between joints and forces governed by gait biomechanics, as well as the relationships between forces and floor responses governed by the structural dynamics. As a result, our model poses physical constraints to reduce uncertainty while allowing information sharing between the body and the floor to make more accurate predictions. We evaluate our approach with 20 participants through a real-world walking experiment. We achieved an average of 3.7 degrees of mean absolute error in estimating 12 joint flexion angles (38% error reduction from baseline), which is comparable to the performance of cameras and wearables in current medical practices.

Figures (5)

• Figure 1: Conceptual diagram of our new HSI system.
• Figure 2: Diagram of our physics-informed graph (PIG) describing the relationships between hip, knee, ankle joint motions, floor vibrations, gait cycle time, and body properties. The nodes of the graph are represented as solid circles with various colors. The edges are represented as arrows.
• Figure 3: Information flow between the vibration nodes and force nodes to enforce structural dynamics equation.
• Figure 4: Experiment setup with geophones and Motion Capture (MoCap) cameras for real-world walking.
• Figure 5: Results comparison between baseline model (LSTM) and our physics-informed graphical (PIG) model.