SenseAgent: An LLM Agent for Adaptive Cross-Domain IMU Sensing

2026-09-25

AI Summary

This study addresses the significant performance degradation of IMU-based activity recognition across users, devices, or postures, where conventional static adaptation struggles with dynamic domain shifts. To overcome this limitation, this work proposes a closed-loop adaptive framework driven by large language model (LLM) agents. It transforms fixed inference pipelines into a “diagnose-plan-execute” loop, leveraging source-domain experience and online target memory to enable prototype transfer. Furthermore, a multi-stage verification mechanism is introduced to reject high-risk decisions, facilitating retraining-free adaptation under scarce feedback. Evaluations across multiple datasets demonstrate substantial improvements in cross-domain sensing performance, particularly in complex shift scenarios when integrated with minimal user feedback. This research establishes a novel paradigm for generalized deployment in ubiquitous sensing applications.

Abstract

Deep learning has improved inertial measurement unit (IMU) sensing for mobile and wearable applications. However, an IMU model trained in one domain often becomes unreliable when it is used with a new user, device, or body position. Existing methods usually treat this problem as a static model-design task: they pretrain a stronger representation, add data augmentation, or select one adaptation method before deployment. In practice, the target domain is only gradually observed, labels are scarce, and different domain shifts require different sensing actions. This paper presents SenseAgent, an LLM-guided sensing agent for cross-domain IMU activity recognition. Instead of asking an LLM to classify raw IMU signals, SenseAgent uses the LLM as a runtime planner over sensing tools, source-domain experience memory, online target memory, and verifiers. The agent builds a label-free diagnosis report from the target stream and uses it to decide whether to keep raw inference or invoke specialized tools, including gravity-aware sensing, prototype transfer, and style normalization. Verifiers check source calibration, target-memory reliability, and no-harm criteria before accepting high-risk tool decisions. SenseAgent also supports scarce feedback without retraining the backbone or replacing the label-free route. This design converts cross-domain IMU sensing from a fixed inference pipeline into a closed-loop sensing process that diagnoses target shifts, selects suitable sensing actions, and rejects unsafe adaptations. We evaluate SenseAgent across multiple IMU datasets and deployment shifts. Results show that its verified route selection improves cross-domain sensing, especially under harder placement and compound shifts, and further benefits from limited user feedback.

Problem

• Cross-domain IMU sensing
• Domain adaptation
• Activity recognition
• Distribution shift
• Wearable computing

Innovation

• LLM Agent
• Cross-Domain IMU Sensing
• Closed-Loop Adaptation
• Runtime Planning
• Safe Verification