SPARQEE an Autonomous Robot

Pre-Competition

Competition Day

Results

Unfortunately, during the competition day SPARQEE did not perform as expected. During the first round one of our H-bridges got unplugged and by the time we figured out that was the issue we were out of time. Once we got that fixed and checked all the other connectors we thought we were ready for round two and were pretty confident we would at least score 2 points and if we got lucky - 6. Unfortunately SPARQEE didn’t agree with us and we ended up scoring no points at all, although it wasn’t due to not being able to move this time.

In second round we tape-followed very smoothly to first Ewok, but SPARQEE failed to identify it, and decided to abandon it. Once it abandoned the first Ewok, SPARQEE crossed the gap and tried to identify the second Ewok but failed. We knew that Ewok was detected for split second in each case, since SPARQEE would stop right in front of it, but it would stop detecting Ewoks shortly after. We kept restarting the course until we ran out of time, but each time the same thing would happen.

Upon thorough analysis of the images that were used to detect Ewoks during competition day we noticed that the contrast between the surface and the Ewok wasn't as high as in our test environment, and that was due to uneven lighting in the lecture room during the competition day (front lights were off in order for the audience to see the projector screens).

Takeaways

The key takeaways of this course and the competition can be split into two main categories: technical and project management.

As a team, we were fairly happy with the way our mechanical and electrical design turned out to be, however there were still avenues for improvement that could have prevented some of the failures during the competition day.

1. More robust methods of connecting important electrical components, such as motors and h-bridges
2. Keeping important electrical components easily accessible, however also protected enough from other moving parts of the assembly. As an example, our basket lifting mechanism somewhat interfered with the internal wiring and might’ve been a cause for the disconnected h-bridge in round one.
3. Due to going for modular design with our circuitry we sacrificed compactability, which resulted in our robot having larger chassis and therefore being not as maneuverable.
4. Certain parameters in our code, such as gap crossing speed, ramp climbing speed, and stopping time, would change drastically as the battery voltage decrease. This caused some indeterminacy in our robot’s behaviour. In hindsight there were a few things that we could have done - downregulate voltage of the main battery to the acceptable value that we new the battery would never fall down to, or, dynamically monitor the charge of the battery and adjust the the necessary values by some power multiplier.

There were two main strategies used for tackling this project:

1. Each person in a group was assigned a specific major component of the robot based on their strength - either mechanical design, electrical design, software design, or manufacturing.
2. Each person was assigned many different tasks from various major components, such that each group member got to do some mechanical design, some software, some hardware, and some manufacturing.