Apple Pi Robotics
Swerve drive is a type of drivetrain where each wheel can rotate independently which allows for full holonomic motion. Previous designs had little to no documentation or testing procedures, and numerous failure points.
When I took over the design process I started with an FMEA on my team’s prior design to understand where it was failing and how I could improve. One of the main problems with the design was how slow it was to manufacture. The module was also significantly bulkier and heavier than it needed to be. Screws in tapped holes repeatedly came loose, the wheel diameter was very small which led to rapid wear, and use of chains to drive the wheel created significant backlash.
My first goal in designing the new modules was reducing the number of operations required to manufacture the modules. I replaced one of the bushings with a 3D printed part and the other bushing could be manufactured in house rather than relying on a sponsor. I removed over a dozen screws from the design, and replaced the majority of tapped holes with nylocks.
To expedite the assembly and machining process I created a detailed machining and assembly guide. I highlighted critical parts of the manufacturing process, and outlined tests the machinist could perform to ensure the part is manufactured and assembled correctly.
I increased the wheel size from 2” to 4”, which allowed for a lower overall gear ratio and less wear on the wheels. I also replaced the module from gear and chain driven to belt driven which almost entirely eliminated backlash.
I wanted to make sure the knowledge I gained from redesigning the modules was not lost when I left the team. So I mentored younger students in the design and involved them in the assembly process. Because of my success, I was invited to present at the regional and world level to over 600 people on my design process.
I designed a climbing system that enabled a 150lb to climb onto a 2’ platform in under 10 seconds. The climber consisted of a single stage continuous elevator that extended below the frame of the robot. The climber had powered wheels on the bottom which propelled the robot forward onto the platform.
The climber had to be fully designed and constructed in under 4 weeks. The shortened time frame and limited machining resources of the team factored heavily into my design decision. All of the components can be made on a manual mill or band saw, and a large number of off the shelf parts were used to expedite the process.
In competition the climber had a failure rate of less than 1 in 15 and was one of the most reliable climbers in the district.