BenchHAR: benchmarking self-supervised learning for generalizable sensor-based activity recognition

Unified framework for evaluating how SSL methods generalise for sensor HAR on unseen target distributions. Curates a ~258K-sample dataset and evaluates 8 SSL methods × 12 encoder–classifier architectures. Findings: existing SSL generalises poorly; hybrid reconstruction + contrastive pretraining is best overall; CNN encoders learn the most generalisable representations, richer classifiers help further; more pretraining data from downstream activity classes helps consistently while more labels add little.

• Form factor: multi

Human Activity Recognition (HAR) from wearable sensors supports broad healthcare and behavior science applications. However, data heterogeneity and the scarcity of labeled data limit its real-world generalization. Recent advances in self-supervised learning (SSL) in vision and language domains have shown strong capability for learning generalizable representations from unlabeled data. Yet, few studies have systematically compared the generalization performance of SSL methods or explored how to adapt them for generalizable HAR. To address these gaps, we present BenchHAR, a unified framework for evaluating the generalization capability of SSL methods for sensor-based HAR on unseen target distributions. BenchHAR curates a large-scale dataset (~258K samples) and evaluates eight representative SSL methods across 12 encoder-classifier architectures. Our results reveal that existing SSL methods struggle to achieve satisfactory generalization performance. We find that: (1) For HAR models, the hybrid paradigm (combining reconstruction and contrastive pretraining) achieves the best overall performance. The CNN encoder exhibits the strongest ability to learn generalizable representations, while more expressive classifier architectures further improve generalization. (2) For data scale, increasing the amount of pretraining data from downstream activity classes consistently improves generalization, while adding more labeled data yields limited gains. Interestingly, incorporating unlabeled data from non-downstream activity classes does not improve generalization. (3) Sensor data collected from custom-grade devices generalizes better than that from research-grade devices, and data from limb transfers more effectively to trunk positions. BenchHAR provides a unified benchmark and actionable insights for generalizable sensor-based HAR systems. Our code is available at https://github.com/saiketa/HAR-Bench.

Verbatim from DataCite (opens in a new tab) · Licence: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

No results from this paper on ChatPPG leaderboards yet. Browse the leaderboards.