TinyHAR-UQ: Battery-aware, uncertainty-controlled tinyML for wearable activity recognition on IoT edge devices

Key Points

• Enhanced activity recognition using tinyML improves performance on IoT edge devices, particularly for wearables.
• The primary metric shows a significant increase in accuracy while managing battery usage effectively.
• Assessment using computational models highlights the balance between performance and energy efficiency in smart devices.
• This suggests that energy optimization is crucial for real-world applications, with potential for broader adoption.