MoPFormer: Motion-Primitive Transformer for Wearable-Sensor Activity Recognition

TL;DR

MoPFormer is a self-supervised Transformer-based framework that tokenizes wearable sensor signals into meaningful motion primitives, improving interpretability and cross-dataset generalization in human activity recognition.

Contribution

It introduces a novel motion-primitive tokenization and a self-supervised training scheme for enhanced interpretability and robustness in HAR models.

Findings

Outperforms state-of-the-art HAR methods on six benchmarks.

Demonstrates strong cross-dataset generalization.

Motion primitives improve interpretability and consistency across datasets.

Abstract

Human Activity Recognition (HAR) with wearable sensors is challenged by limited interpretability, which significantly impacts cross-dataset generalization. To address this challenge, we propose Motion-Primitive Transformer (MoPFormer), a novel self-supervised framework that enhances interpretability by tokenizing inertial measurement unit signals into semantically meaningful motion primitives and leverages a Transformer architecture to learn rich temporal representations. MoPFormer comprises two stages. The first stage is to partition multi-channel sensor streams into short segments and quantize them into discrete ``motion primitive'' codewords, while the second stage enriches those tokenized sequences through a context-aware embedding module and then processes them with a Transformer encoder. The proposed MoPFormer can be pre-trained using a masked motion-modeling objective that reconstructs missing primitives, enabling it to develop robust representations across diverse sensor configurations. Experiments on six HAR benchmarks demonstrate that MoPFormer not only outperforms state-of-the-art methods but also successfully generalizes across multiple datasets. More importantly, the learned motion primitives significantly enhance both interpretability and cross-dataset performance by capturing fundamental movement patterns that remain consistent across similar activities, regardless of dataset origin.