cancel
Showing results for 
Search instead for 
Did you mean: 

Fall detection not working with STM32F401RE & IKS01A2

EPark
Associate

Hi

I ​ expected & need to detect human(senior or kid) falling to ground not free falling.

When I tested with the pre-built binary named "STM32CubeExpansion_MEMS1_V6.2.0\Projects\STM32F401RE-Nucleo\Applications\IKS01A2\FallDetection\BinaryFallDetection.bin" with STM32F401RE & IKS01A2 board, The free falling detection function was worked sometimes but human falling was never detected.

I tried several times with various situation of falling action. I couldn't find the detect condition of human falling.

your fall detection algorithm works for only free falling ?

I want to make a product of senior watch solution for detecting fall and sending alarm message to pre-defined customer's family.

please let me know the human falling detection work or not.

Thanks

2 REPLIES 2
Eleon BORLINI
ST Employee

Hi @EPark​ , the Motion FA algorithm should detect not only the free fall, but also the "guided" fall of the board + the falling object according to UM2275. The fact you are not able to detect human fall could be related to a miss in the free fall threshold detection. The falling object and the STM32F401RE & IKS01A2 board have to be as tied together as much as possible: consider also that the system you are using is quite big for your target application... How did you tested it? Regards

EPark
Associate

​I tested with STM32F401RE & IKS01A2 & IKS01A3. The EV board connected to my PC by USB cable. I dropped the EV board to floor with cushioning for protecting from damage. When the EV board turned over after dropping, fall was not detected always.

The result so far tested, It is not suitable for my application(senior watch). I think that gyro sensor data can be helpful  for more accurate fall detection.

I cannot see and modify the source code of the MEMS fall detection library. but I need to modify it.

Please let me know the contact point of 3rd party company who developed the MEMS fall detection library.

Thanks