search

ProjectNeura / LEADS

Enable your racing car with powerful, data-driven instrumentation, control, and analysis systems, all wrapped up in a gorgeous look.
https://leads.projectneura.org
Apache License 2.0
251 stars 53 forks source link

Improve the Tolerance of the Wheel Speed Algorithm #178

Closed ATATC closed 5 months ago

ATATC commented 5 months ago

As it was noted during today's test #171, the wheel speed sensor frequently failed to read a value due to the shaking of the wheels. Therefore, we want to improve the tolerance of the algorithm so that it does not compromise easily.

Also, it has been brought to attention that although the vehicle was at rest, the program still outputs some random speeds. Further investigation is required.

ATATC commented 5 months ago

A Multi-sensor Hypothetical Fusion Inference Algorithm

When there is a huge difference between the readings of different wheel speed sensors, we assert that there is failure. We use other available valid data to infer the speed. If multiple inference suggest speeds that fit in an acceptable range, take the smallest reading the lies in that range as the final output.

Definition of "huge": the difference is greater than the overall variance of the speed sequence.

Implementation

The implementation of this algorithm deeply depends on the utilities introduced in #173. I'm going to make the inferences available without matplotlib. Little bit concerned about the performance, but should be fine.

@qmascarenhas Please review.

ATATC commented 5 months ago

@qmascarenhas This is way too hard and time-consuming to realize. I won't be able to finish it within a week. I propose we go with better mounting methods.

qmascarenhas commented 5 months ago

No problem @ATATC - it shouldn't be too difficult to find a way to reliably secure the magnet to the wheel spoke without it twisting due to wheel/spoke vibration. We can postpone this task to after the race.

ATATC commented 5 months ago

In my investigation of the zero speed problem (replaying mode really helped), I suspect that it might be caused by the "enhancement" of the wheel speed algorithm introduced 4 months earlier. I'll try to fix that first. I suggest we schedule another test next week before the race.

qmascarenhas commented 5 months ago

We can test all we want on the temporary cart setup we have. We need to assign priority to the test drives this week to get the drivers and pit crew race ready. Our work on and with VeC99 is going to have to happen outside of their time with the vehicle until we can assure them that we have a stable solution that will not disrupt their practice routine. We need to be able to race VeC99 without LEADS so if there is a LEADS catastrophic failure on race day there is no panic.

ATATC commented 5 months ago

Do we still have enough time to install the instrumentation system?

qmascarenhas commented 5 months ago

Yes! I need to do one more test on the cart with the longer wheel speed sensor wires (those measurements you took on Friday) and we should be good. If you could give me a hand with that this week while I iron out some of the mechanical installation issues, that would be great!

ATATC commented 5 months ago

Yes sure. I should be available any afternoon this week.

ATATC commented 5 months ago

Wednesday is off. Other days are free.

qmascarenhas commented 5 months ago

ok. Let's try to work to the following schedule then:

ATATC commented 5 months ago

Do we have a time schedule for the race day?

qmascarenhas commented 5 months ago

https://uwaterloo.ca/electric-vehicle-challenge/races/waterloo-ev-challenge

8:30 am - 9:15 am | Teams arriveRegistration opens 9:15 am - 10:15 am | Safety inspections, award judgingTransponder setup and inspection 10:15 am | Driver's meeting in pits -- mandatory attendance by the 12 and 24 volt drivers 10:30 am | Opening ceremonies, all 12 volt cars on track 10:45 am - 12:00 pm | 12 volt race  (60 minute duration) 12:00 pm - 1:30 pm | Lunch break and prep for 24 volt race 1:30 pm - 2:45 pm | 24 volt race (70 minute duration) 3:00 pm | Awards ceremonies The time allotted to each race is longer than the official race time — this is to allow for cars to be stopped, moved off the track, repaired.

ATATC commented 5 months ago

Can't see it.

qmascarenhas commented 5 months ago

i just updated the comment.

ATATC commented 5 months ago

Did u tell Jerry that he needs to come to my place at 7?

qmascarenhas commented 5 months ago

No I didn't. Can you sort that out with him directly?