For a research project (human body motion), I am looking for a (low-cost) open-source (hardware/software) wearable device for fast ML prototyping (test ML-models in lab situations i.e. inference on the device).
Programmable with tools like PlatformIO and/or mbed.
Small size, 50mm x 40mm x 20mm, enclosure/case included, MCU (M4) + BLE link (for communication with laptop) + battery and 9-DOF IMU or free space/extension for an (extra) IMU, example MTi-1 IMU).
Any suggestions?
The idea is to have a minimum number of components, only MCU, IMU, storage (for IMU data) and BLE, no extra sensors (such as microphones etc) and/or extra components.
Ideal, something like this: Xsens DOT, but completely open source.
Currently, I am playing around with SparkFun MicroMod Machine Learning Carrier Board on a SparkFun MicroMod Machine Learning Carrier Board in combination with SparkFun 9DoF IMU Breakout - ICM-20948. This to explore what is possible.
Currently, one of the ideas is to redesign the carrier board (remove the components which are not needed) and add MTi-1 IMU and try to fit it in OKW Body case (M, L or XL), OKW Body Case.
However, I am exploring if there is something available, open-source, so I can short-cut some design time.
The sensor will be used in academic research and education: embedded design and tinyML.
It is 50mm diameter which is 10mm bigger than your requirement and only has a 3dof IMU onboard but Iāve had some successes with running tinyml models on it. But of note, it does have an nrf52840 chip which is the same as the Arduino nano 33 BLE sense and a built-in JST lithium polymer battery adaptor. The circuit playground are popular in hobby/costume wearable tech projects, so I highly recommend it if you need an alternative to the arduino nano 33 BLE sense.
p.s. Iām not sponsored by either Arduino or Adafruit (or anyone!)ā¦ just an enthusiast whose used their products .
The idea is to have a minimum number of components, only MCU, IMU, storage (for IMU data) and BLE, no extra sensors (such as microphones etc) and/or extra components.
Ideal, something like this: Xsens DOT, but completely open source.
Currently, I am playing around with SparkFun MicroMod Machine Learning Carrier Board on a SparkFun MicroMod Machine Learning Carrier Board in combination with SparkFun 9DoF IMU Breakout - ICM-20948. This to explore what is possible.
Currently, one of the ideas is to redesign the carrier board (remove the components which are not needed) and add MTi-1 IMU and try to fit it in OKW Body case (M, L or XL), OKW Body Case.
However, I am exploring if there is something available, open-source, so I can short-cut some design time.
Not sure about āfully open sourceā, I believe that the Arduino core might be not open source on this one, but Iām not sure. It exists in more than one version: apart from ESP32 thereās also one with nRF52832 MCU, and also the IMU varies (on some models itās a TDK chip, not STMicro).
.