cancel
Showing results for 
Search instead for 
Did you mean: 

how to enable or disable tampering detection on LIS2DH12TR sensor? how to validate whether the interrupt is generated by tampering detection?

Mnago.1
Associate II
 
1 ACCEPTED SOLUTION

Accepted Solutions
Eleon BORLINI
ST Employee

Hi @Mnago.1​ ,

What do you mean with "tampering detection"? Would you like to be able to detect just an impulsive crossing of an acceleration threshold, or also a specific pattern (corresponding to a "tampering" action)?

In the first case, you could use the "wake-up" feature and interrupt. Suppose for example that you place the LIS2DH12TR inside a safe: when the theft try to open it, he will force it, generating some spikes. But a simple threshold crossing would be hardly distinguished from a correct (and "soft") opening pf the safe.

For this reason, the second option is to use a device (accelerometer and/or gyroscope) with the capability to detect more complex patterns, for example using features like the finite state machine (FSM) or, maybe better, the machine learning core (MLC) such as for example the LSM6DSOX. With this last feature you can create a reference dataset of the movement you want to detect and gain an accurate result on the tampering motion detection (a tutorial for its usage can be found here).

I see you also wrote this other post, so please refer to this one for the other topics.

-Eleon

View solution in original post

5 REPLIES 5
Eleon BORLINI
ST Employee

Hi @Mnago.1​ ,

What do you mean with "tampering detection"? Would you like to be able to detect just an impulsive crossing of an acceleration threshold, or also a specific pattern (corresponding to a "tampering" action)?

In the first case, you could use the "wake-up" feature and interrupt. Suppose for example that you place the LIS2DH12TR inside a safe: when the theft try to open it, he will force it, generating some spikes. But a simple threshold crossing would be hardly distinguished from a correct (and "soft") opening pf the safe.

For this reason, the second option is to use a device (accelerometer and/or gyroscope) with the capability to detect more complex patterns, for example using features like the finite state machine (FSM) or, maybe better, the machine learning core (MLC) such as for example the LSM6DSOX. With this last feature you can create a reference dataset of the movement you want to detect and gain an accurate result on the tampering motion detection (a tutorial for its usage can be found here).

I see you also wrote this other post, so please refer to this one for the other topics.

-Eleon

Eleon BORLINI
ST Employee

Hi @Mnago.1​ ,

do you have any updates on your application?

-Eleon

Mnago.1
Associate II

the given first case above is the one my application implements.

and does it works?

-Eleon

Mnago.1
Associate II

hi, yes the tampering detection works, as i am configuring sensor to higher threshold for detecting the tampering event.