Rapid system streamlines machine learning model deployment on wearable devices
A Rapid Pipeline for Training and Deploying ML Models on WeBe Band
Artificial IntelligenceMachine LearningPerformance
Summary
Making machine learning models work well on small devices like wristbands is hard because these devices have limited memory and power. The authors built a system that helps create, test, and install machine learning models quickly on the WeBe Band, a wearable that monitors health data. Their system automatically makes models that fit the device’s limits and can be updated wirelessly. It lets developers try different kinds of models and see how fast and efficient they run on the actual device. This work focuses on making the process easier and faster rather than inventing new types of models.
What this means in practice
- •For wearable device developers: Develop and deploy customized machine learning models rapidly on wrist-worn health monitors with automated integration and hardware-aware optimizations.
- •For embedded systems engineers: Profile and optimize various lightweight machine learning models on constrained microcontroller platforms for real-time physiological data processing.
- •For medical device manufacturers: Create and update intelligent health-monitoring firmware on wearable products through wireless deployment of efficient ML models tailored to device constraints.$Commercial implications: Enables selling wearable health products with regularly updated intelligent features optimized for low-power hardware.
Authors
Ehsan Kourkchi, Asmita Asmita, Houman Homayoun, Mahdi Eslamimehr
Abstract
Developing optimized machine-learning algorithms for edge devices with limited computational and memory resources is challenging, time-consuming, and highly dependent on device-specific constraints. In this work, we streamline an edge ML workflow to enable rapid development, optimization, and deployment of machine-learning (ML) models directly on the WeBe Band, a wrist-worn wearable device designed for multimodal physiological data monitoring. The proposed system automatically generates hardware-efficient ML models that can be easily integrated into the WeBe core firmware, supporting AutoML, hardware-aware quantization, and performance profiling to build models that meet desired latency targets while remaining compatible with device memory and power limitations. The proposed framework tightly integrates the open-source Piccolo AI ecosystem with an automated pipeline that generates deployable firmware artifacts, performs hardware-aware model compilation, and supports over-the-air (OTA) deployment. The system supports multiple lightweight model classes, including classical machine-learning algorithms and neural networks, and provides built-in on-device profiling tools to evaluate inference latency and memory footprint under realistic execution conditions. Experimental results demonstrate clear trade-offs between model complexity and deployability on a microcontroller, showing that classical models offer strong real-time performance while lightweight neural networks require careful resource management. Rather than proposing new learning architectures, the current work mainly focuses on system-level automation, deployability, and enabling researchers and developers to rapidly iterate on models and evaluate them directly on target hardware. Although demonstrated on the WeBe Band platform, the workflow is designed to be extensible to other ML-powered edge devices.