Project Background

For a connected motion tracking device, raw sensor data must be captured accurately, filtered, and processed into reliable motion data before it can be used for meaningful activity analysis.

This was the foundation of a wearable technology platform developed as part of an Innovation Fund project. The goal was to create a flexible platform for wearable devices that could collect motion sensor data, perform signal pre-processing and sensor fusion in firmware, and generate 10-axis motion data. This processed motion data could then be used by AI algorithms to analyze movement patterns based on predefined motion models, supporting applications in sports, fitness, workforce monitoring, and equipment usage analysis.

In practice, this means the technology can provide feedback on exercise technique and performance, support injury prevention by identifying risky movement patterns, and help equipment manufacturers understand how their products are used in real conditions.

The Challenge

The main engineering challenge was developing a compact embedded device capable of collecting motion data, pre-processing sensor signals, and supporting AI-based motion pattern recognition directly on the device.

Raw motion sensor data is often noisy, fragmented, and difficult to use without proper processing. To make it useful for motion analysis, the firmware had to support sensor fusion and pre-processing algorithms that could generate reliable 10-axis motion data for further interpretation.

An additional challenge was the implementation of wireless charging and local flash storage into the wearable platform. Both functions had to be implemented within limited space and power constraints while maintaining stable device behavior and reliable motion data collection.

Our Contribution

Our team supported the development across electronics, firmware, sensor data processing, and mechanical integration.

Electronics design and PCBA verification

The electronics design included the development of a custom printed circuit board for collecting and pre-processing data from motion sensors. The design was created to support embedded motion analysis, sensor fusion, and communication between the hardware and firmware layers. We also provided support during PCB assembly, followed by PCBA design verification.

Firmware development and Zephyr RTOS porting

Firmware was developed in embedded C, and Zephyr RTOS was ported to the device platform. This provided a structured firmware foundation for managing device operation, sensor data acquisition, timing, and future software development.

Sensor fusion and edge AI implementation

We implemented algorithms for sensor pre-processing and fusion, enabling the device to generate 10-axis motion data. Edge machine learning algorithms were also implemented to support motion pattern recognition based on predefined motion models.

Mechanical integration and user interface

On the mechanical and user-interface side, a commercial off-the-shelf enclosure was selected and mechanically adapted to meet the product requirements. We designed the device interface, including the front panel with membrane buttons and a flexible printed circuit, enabling reliable physical user interaction within the selected enclosure.

EMC pre-compliance testing

As the development moved closer to prototype validation, we performed EMC pre-compliance testing to identify potential electromagnetic compatibility risks early and support a clearer path toward further product validation.

Results

The client received a functional platform for motion tracking and edge-based movement analysis.

The delivered platform gave the client:

• a tangible device for demonstrating the MVP to users, partners, and investors
• a functional solution for validating motion tracking use cases in sports, fitness, workforce monitoring, and equipment usage analysis
• edge-based motion analysis capability for recognizing movement patterns directly on the device
• technology platform for wearable products that require motion sensing capabilities