Sensor Technology Research Centre

Research: platforms

We develop our in-house embedded sensing and embedded computing platforms for a number of projects. The advantage of in-house platforms are:

  • Rapid adaptation of the blueprints (hardware, firmware) to new research projects
  • Exploiting commonalities across research activities
  • Exploration of new sensors for behaviour analytics
  • Exploration of new computing for behaviour analytics (e.g. multicore, CPU+FPGA, dedicated silicon)

Currently two main platforms are available: BlueSense2 (8-bit AVR) and BlueSense4 (32-bit Cortex M4).

BlueSense2 - Wearable/IoT sensing platform and inertial measurement unit



BlueSense2 is a wearable/IoT platform designed to be functional out-of-the-box yet extensible. It's primary purpose is to be an inertial measurement unit for wearable applications. Yet it is extensible for IoT applications with a number of features making it appealing for this purpose

BlueSense2 as a wearable platform is a tiny device which can be used to capture body movement with a 9DoF motion sensor. The data can either be logged on an SDCard or streamed over bluetooth. This forms the core of a fitness tracker. Its extension ports allow to plug-in additional sensor modalities for research purposes (e.g. electric potential sensors). A display is being developed converting it into a smartwatch.

As an IoT device it's main appeal is a true hardware off which allows to put the device entirely to sleep (everything is powered down, including voltage regulators), yet wake up at programmed intervals thanks to a real-time clock wakeup. This allows to achieve very long battery life, as the device can wake up at desired times (once an hour, once a day, ...) to acquire and send sensor data. The extension ports allow to plug in custom sensor modalities which make this device highly versatile.


  • ATmega1284P core (128KB flash, 16KB RAM)
  • 30x30mm
  • USB interface for charging, device interaction
  • Bluetooth interface for device interaction
  • SD card (SDHC standard)
  • Most accurate RTC (real-time clock) on the market: +/-5ppm
  • 9DoF motion sensor (MPU9250) including software attitude and heading reference system at 100Hz (quaternion output). The system is capable of acquiring and storing raw motion data at 1KHz, which is useful for high frequency applications (sports, vibration analysis)
  • Extension ports with analog inputs, digital I/O, I2C, SPI, timekeeping, etc.
  • Coulomb counter allowing precise measurement of battery charge/discharge and characterisation of real-time current consumption by software


Roggen et al. BlueSense: designing an extensible platform for wearable motion sensing, sensor research and IoT applications, Proc. International Conference on Embedded Wireless Systems and Networks, 2018

Ciliberto et al. Complex human gestures encoding from wearable inertial sensors for activity recognition, Proc. International Conference on Embedded Wireless Systems and Networks, 2018

Roggen et al. Electric field phase sensing for wearable orientation and localisation applications, Proc. 2016 ACM International Symposium on Wearable Computers, 52-53, 2016

Pour Yazdan et al. Wearable electric potential sensing: a new modality sensing hair touch and restless leg movement, Proc. 2016 ACM International Joint Conference on Pervasive and Ubiquitous Computing: Adjunct, 846-850, 2016


Version Release date Changes
9 October 2017 Changed the MPU interrupts to exploit hardware timers for downsampling
8 June 2017 Fixed the RTC wake-up logic
7 November 2016 More I/O on expansion port. First production version.
6 May 2016 Improved USB interface
5 Jan 2016 Improved motion sensor acquisition speed (1KHz)
1-4 2013-2015 Initial prototypes

Documentation & source

BlueSense4 - Wearable/IoT sensing platform and inertial measurement unit

BlueSense4 is major redesign of BlueSense2 using a state of the art ARM Cortex M4 processor (STM32L496) with advanced low-power functionalities, an updated motion sensor (ICM-20948), a built-in digital microphone, and extended expansion connector. The rest of the hardware features are similar to BlueSense2.

The firmware user interface is compatible with BlueSense2, allowing somebody familiar with BlueSense2 to use BlueSense4 immediately. Additional functionalities of the firmware include:

  • Sound acquisition
  • Multimodal acquisition: motion, ADC inputs, sound
  • DAC signal generation (multiple mixed signals with individual frequencies and amplitudes)
  • Enhanced ADC acquisition (12 bit with a compact binary frame format for high speed acquisition)

Documentation & source


Version Release date Changes
3 October 2019 Changed MPU to ICM20948; changed microcontroller to STM32L496
2 June 2019 Upgrade to feature parity with BlueSense2
1 October 2014 Initial version with MPU9250 and STM32F401 microcontroller