cancel
Showing results for 
Search instead for 
Did you mean: 

Air Gesture Recognition using IMU

Pointes
Associate II

I want to implement air gesture recognition using accelerometer and gyroscope data. 
With different movements you should be able to make settings on the handheld device. 
Movements to be recognized are, for example, tipping twice or moving up and down twice

My findings so far:

  • The first attempt was made with LSM6DSO and it's Finite State Machine (FSM) by using Unico tool. Basically it works, but the reliability is too low. 
  • The second attempt was made with LSM6DSOX and it's Machine Learning Core (MLC) by using Unico tool. Reliability is too low in this case as well.
  • The third attempt was made with Qeexo AutoML using the SensorTile Box MCU for processing the sensor data. Similar to this description. https://community.st.com/t5/mems-and-sensors/how-to-detect-hand-air-gestures-with-the-sensortile-box-and/ta-p/49547
    The result was relatively good. However, just one tipping triggers the double tipping event. 

My biggest problem is, that I don't want to detect continuous signals, but specific time discrete events. I had Qeexo AutoML trained with segmented signals. Nevertheless, it triggers with just one tilt and not with two tilts. Apparently there are difficulties with the time window to be examined.

My next consideration is to find a possible solution with NanoEdge AI or STM32 Cube.AI.
Can anyone tell me what is the best solution to reliably detect time discrete events using 6 DOF accelerometer and gyroscope data?

13 REPLIES 13
Pointes
Associate II

Hi @niccolò ,

this would be a workaround for some gestures.

Can another sensor recognize these gestures that would be better for my application?
Maybe the LSM6DSO16IS with its ISPU - intelligent sensor processing unit?

Pointes_0-1698946773575.png

It is mentioned that it can even do digital gesture recognition.

Pointes_1-1698946898792.png

 

 

Hi @Pointes ,

you can try to use the ISPU to get to better results, but the digit gesture recognition is linked to the Qvar, so you should add a sensor embedding the Qvar.

anyway, with the ISPU you have a lot a flexibility in developing your own neural network but you have to be an expert in C programming to be able to implement it.
automated tools to develop networks are not available now.

Niccolò

p.s.: if you'd like more info on this matter, feel free to open a new thread, so that even the other users can follow =)

pratham
Associate

Can you give me some more details about your project actually i am also having the same plan of project but with different hardwares and softwares.

Hi @pratham 
Since I posted the question here in the forum, I have not continued with the project. But I will continue at the end of this year.
Unfortunately, the LSM6DSOX can't do some basic functions that are essential for me. Therefore, I will probably do the calculations on the microcontroller and only request raw data from the sensor. I don't yet know exactly what this will look like.