PiWars 2019: ... In The Beginning
Yesterday we held our first PiWars 2019 meeting. Many many things were discussed including our past performances at PiWars 2017 and 2018, some of the problems we had that we still need to overcome and … most importantly … what are our plans for the next 6 months to make PiWars 2019 a raging success for Maidstone Hackspace.
Autopsy results for PiWars 2018:
- Control (lack of)
- Omni wheels allow us to turn with little torque, but don’t stop in time (read: power sliding everywhere)
- Ramping up/down on the motor driver boards possibly a cause? Try without.
- 3d printed chassis – layer adhesion issues
- Letting Ian drive it
- Programming for minimal maze
- 1 successful run though.
- ToF sensors lost connection.
- No PID to keep distance from wall stable.
- Programming for speed test (no PID)
- This was an all out flop which wrote off the mounts for the ToF sensors.
- We had no time to even test this before the event.
- Programming over the rainbow
- Too basic movement controls (overshooting the targets and no logic to correct itself)
- Didn’t pre-seek the corners.
- Again, relatively untested beforehand.
Proposals for success
- New chassis (aluminium) – Oly’s CNC
- Captive nuts in motor brackets rather than in the chassis (as it won’t be 3d printed this time)
- OpenCV for speed test (line following) and over the rainbow.
- Mirror ball camera for 360 view of over the rainbow.
- Time of flight for maze
- Mecanum wheels (3d printed)
- PiNoon macros
- Jake must practice driving!
- Blogging - after every meeting - we’re going to be really open - no big secrets
- A body (within the Apollo theme)
- Tidier wiring
A note about how things will be…
The PiWars meetings will take place on a day other than the Wednesday Hackspace open day. You can still contribute if you can only come on a Wednesday, but it will give the project a dedicated time to move forward.