Analog stick issue with iCubGenova04 joypad

[paolo-viceconte]

There was the problem of the joypad sending unexpected rotational motion commands to the WalkingModule as soon as we started the controller, even when the analog stick was not touched. This persisted even after restarting the joypad modul.

We thought possibly it was related to deadband issue. But with a bit of debugging with prints from the joypad module, we noticed that the iCubGenova04 joypad sends non-zero values for the y axis of the left analog even if no one is touching the analog. From time to time the sent value is far away from zero (like bigger than 1). This seems to be the source of the occasional turning on the spot. The non-zero values are resetted to zero only when we press the stick.

VID_20210827_151230.1.mp4

So as a workaround for testing walking controller, we tried connecting another joytstick following the tutorial for pairing bluetooth devices with icub-head (in general Linux OS) is provided in https://icub-tech-iit.github.io/documentation/icub_operating_systems/icubos/bluetooth/.

We ran bluetoothctl.

There was already Joypad device SteelSeries with address 28::9A::4B::08::EC::67 (iCubGenova04 joypad) registered as a trusted/paired device.
We did a scan of the devices using scan on.
We registered the iCubGenova09 Joypad with address 28::9A::4B::08::E1::1B as a trusted device as well. This was done by running commands,

scan off
trust 28::9A::4B::08::E1::1B
connect 28::9A::4B::08::E1::1B

After, doing so we were able to run walking with iCubGenova09 joypad.

We need to remember to remove the pairing of iCubGenova09 after demo.

cc @prashanthr05 @isorrentino @S-Dafarra @Yeshasvitvs @DanielePucci