Project Logs

Collapse

• Using Edge Impulse with Tracer to build your own custom gesture/activity tracker

  YJ • 3 days ago • 0 comments

  I have been inspired over the last few weeks by Fabio's project with Edge Impulse and how the Tracer platform can be used to leverage the ESP32's processing power to run neural network classifiers of activities and gestures. In this example, I will go through the steps needed to train your own neural network and show it performing live classification with data streamed directly from the board. 

  I wrote an Arduino sketch that would spit out the 3-axis accelerometer and 3-axis gyrometer at a frequency of 100Hz. The experiment I had in mind was basically to use Tracer as a classifier that could tell if it was moving horizontally, vertically, "caterpillar" (move in a jagged way horizontally), and waving. A dataset of 10s was collected for each of the movements.

  Once I had an idea of what I wanted to build, I followed the steps shown in the video tutorial from Edge Impulse in getting a model trained up as well as a k-means-based anomaly detector.
  

  The initial model that was trained showed inaccuracies when it came to classifying a wave and that is due to several reasons but I think it was primarily due to not having enough data. The table below shows the confusion matrix for the first training cycle with the original dataset.

  Following that, I wanted to see if collecting more data would improve so I spent another 10 seconds collecting data for wave, and it improved slightly. I think more data in each category should improve the neural network's specificity and sensitivity. The accuracy improved across the board while the error rates also dropped.

  Here is a short video example of how the live classification tool is being used on the data that is streaming directly from Tracer over USB (UART). The next step for me is to deploy a real-time classifier for tennis shots that will run on the ESP32 and stream the results over BLE.

  And again as before, really happy to hear any feedback, ideas or suggestions!

• Register your interest on Tindie

  YJ • 04/21/2022 at 17:28 • 0 comments

  Hi all,
  
  I am grateful for the warm reception, the development units I had sold out in less than 12 hours and I plan to place an order towards the end of the month. Please register your interest if you are keen on getting hold of this;
  It will ship with the board itself and a high-quality velcro strap that can quickly get you started gaining insights. Due to restrictions on Lithium-Ion batteries, you will have to source the LIR2450 on your own but that is quite commonly available and can be found on Amazon and RS for example. 

  Eager to get these boards out there to see what interesting applications can arise through the open-source community! Also, any feedback is greatly appreciated. 

  Best, 
  

  YJ

• Smaller Tracer designed and manufactured

  YJ • 04/18/2022 at 21:17 • 0 comments

  Hi all,
  
  Just a small update with hopefully bigger ones imminent. I have received my new Tracer boards that are now smaller and have a better-suited LDO with a very low forward drop to increase the usable range of battery voltages from the coin cell battery.
  
  I am currently planning the following activities; 

  1. Testing it out for tennis stroke counting - Done
  2. Tracking bicycle cadence - Done
  3. Using it to synchronize blade strokes at my local rowing club 
  4. Power measurements and optimization - Done
  5. Syncing with camera footage to add data points and improve the usability 
  6. Create a Tracer "hub" to synchronize several modules. Still thinking about the best way to do this as BLE might be a little limiting. Feel free to suggest :) 
  7. Setting up a Crowdsupply campaign to get the hardware out to you! - On Tindie for now

  Thank you all for liking and following this project it means a lot to me and it helps me keep going! 
  Best,

  YJ

View all 3 project logs