Client background
Our client provides advanced technology for smartphones, consumer electronics, and computing platforms. Its portfolio includes biosignal processors that monitor multiple biosignals at once and deliver accurate, real-time data for healthcare and fitness wearables.
Business challenge
Streaming and processing high-frequency biosignal data over low-power wireless channels is hard. Wearables sample continuously, yet they must run for days on a single charge.
The usual approach sends raw sensor data to an edge device such as a smartphone, which processes it and passes it on to the cloud. Real-time feedback for the user drains the smartphone battery quickly, and that limits how practical the solution is.
Project description
To cut battery consumption on end-user devices, SoftServe moved part of the signal processing to the wearable’s DSP. Sending only a small fraction of the data reduces network bandwidth, which in turn extends battery life on both the wearable and the smartphone.
SoftServe’s R&D team built a multi-sensor data acquisition and processing platform that connects wearable, mobile, and cloud services over an Internet of Things (IoT) infrastructure. The team implemented a fog and edge computing architecture using wireless sensors, BLE connectivity, and smartphone and cloud processing. It enables real-time readings of vital signs, other biosignals, and motion data. Data streams from smartphones to cloud storage and dashboards. Android devices handle sensor configuration and real-time signal monitoring, and send data to the cloud via UDP or TCP. Users can access recorded data or live streams from any modern web browser.
The team also implemented the thirty most-used signal processing functions in C for the ARM DSP and tested them on simulators and hardware. Testing was automated end to end, from compilation and flashing to comparing results against reference data. The signal processing algorithms were then updated to use the DSP library. The wearable now handles part of the processing and sends the processed result to the smartphone as a real-time stream.
Business value
Following reference IoT architecture patterns, smartphones act as edge devices. They collect raw sensor data and process time-series data to extract useful features. The data then goes to the cloud for storage and visualization.
The IoT software development kit for biosensors, developed by SoftServe, doubled smartphone battery life and allowed devices to run for several days.
