Rasbian Update Kills Diddy!

Submitted by il diavolo on Fri, 05/04/2018 - 22:30

Forums:

He is still breathing but doesn't respond to PS3 commands. (Diddy is a Mk 1 DiddyBorg with a Pi 3B).

This evening (04/05/18) I did a "sudo apt-get update" followed by "sudo apt-get dist-upgrade". All seemed to go well but when I put the upgraded card back into Diddy I found that most of the commands issued by the PS3 controller were ignored. He still responded to 'backwards' and 'forwards' from the left stick but would not turn left or right and ignored other commands from various buttons.

I reflashed the SD card with the previous image (I always make a back up before upgrading) and Diddy reverted to his normal behaviour indicating that it is not a hardware problem.

I then tried again but this time using "apt-get upgrade" in case some dependencies were being changed but with the same result.

I just wondered if anyone had experienced a similar problem? If there is no solution then Diddy will have to live without being updated which is probably not a good idea.

piborg's picture

piborg

Sat, 05/05/2018 - 13:59

We have not seen that specific issue before, but we have had our fair share of strange connectivity problems when updating in the past.

The good news is that it is likely everything is working fine but the button / axis mapping has changed for some reason.

I would suggest that you use either jstest-gtk (will probably need to be installed by apt-get) or jstest to try and determine the correct axis and button numbers for the PS3 controls.

Once you know the values you can update the "Settings for the joystick" section of the script to match the new numbers. After that everything should be working again like it used to :)

il diavolo

Sat, 05/05/2018 - 21:43

You're right, all the axes and buttons have been remapped (except the left stick axes). Also none of the motion pitch and roll actions seem to be detected any more. This are not a problem at the moment as none of my scripts use them but they were working previously.

As you say, I shall just update the numbers, trouble is I've got lots of scripts for various Diddy tasks and challenges!

By the way, his latest challenge, to replay autonomously a route over which he has been driven manually, is working quite well. I am using both the XloBorg compass and my mouse scroll wheel rotary encoders for distance and bearing of each leg of the route, recording the readings in Python lists and saving them as Pickle data.

il diavolo

Sun, 08/26/2018 - 22:27

The first run of a recorded route works reasonably well, with Diddy arriving quite accurately at the recorded end point at the other end of my living room (about a 5 metre route). There is a problem with getting him to return though, as he usually fails to accurately follow the reversed route. I think the main problem is the pile on the fitted carpet which affects the slip of the wheels during tank steering. However, he doesn't fair too well on the return journey on my tiled kitchen floor either. In this case he passes two large heating radiators and I am wondering if the compass is affected. In both cases the first run would be fairly accurate because the external influences will be the same in both the original recording and the replay but on the return journey the influences will be opposite to those recorded.

It's a pity that GPS is not really usable indoors.

Subscribe to Comments for "Rasbian Update Kills Diddy!"