Last friday, NEUE Labs paid a visit to a noise and vibration testing facility in Stockholm. The reason behind this was to try and identify how much variance we get in the signal from the IMU.

For the people out there that don’t know what variance means in this context, I’ll explain. Think of it as the fluctuation occurring on top of the true values registered by the sensors. For example: say you rotate the sensors at a constant rate of 200 degrees per second (or dps) and the sensor has a variance of 2 dps. That means every time you read the sensor it will give a value between 202 and 198 dps. If you know in advance how much variance a sensor has you can assess how reliable a signal is. This is then used in the algorithms running within AUTOMOTION to determine which signal is giving the correct output.

So we set out to double check how much variance the system has and the only way to do that is to have the sensor perfectly stationary and read the static output. Since the sensors are so sensitive that they can actually pick up background vibration like road vibration we needed a place with very few background vibrations. At the test facility was a room where the entire floor was isolated from the foundation with rubber. In the room itself was a massive vibration rig resting on rubber feet (which we never turned on of course!). These two features combined created a very low vibration environment for us to use. One step closer to a perfect motion sensor and algorithm...

 

Comment