What are the wall sensor used for? Are they worth it?
After looking at the sensor configurations of the winning mice a year or so ago, we decided to follow their lead and implement something similar.
This means two front facing sensors that look directly in front of the mouse. Two diagonal facing sensors that look at the side walls at an angle. Two perpendicular wall facing sensors that look direclty at the side walls.
The thinking is that the front wall and diagonal sensor are used for navigation/correction and the wall sensors are used to determine the presence/absence of walls, and to look for edges for longitudinal correction.
Since then, mice have started eliminating the wall sensors. In fact, the code that Zeetah V (ZV) used at the Japanese contest this year, used only the front and diagonal sensors. Also, due to an accident, two traces for the right wall sensor were damaged.
So, the thought has been, why not eliminate the wall sensors? It turns out that for the ZV sensor configuration, during diagonals, the wall sensor would have been helpful but beyond that one case, there was no other benefit. In fact, using the wall sensor for wall state determination meant that ZV had to go further into the cell before it could make that decision than if we continued to use the diagonal sensor.
The wall sensors are also the furthest element from the center on ZV. So what, you ask? The sensor blocks weigh 1.75gms each (3.5gms total out of 100gms for the mouse) but they contribute 11% of the rotational inertia!
They are now history…