Initial Brainstorming
We started out by reading through the prompt and establishing our initial design requirements. We prioritized the ability to drive, hold balls, and shoot balls, followed by stopping at the fact checker and safe space, and being able to navigate.
After determining our initial requirements, we sketched out designs for some of the main subsystems: the drive system, the shooter, and the line-sensing circuitry
pROTOTYPES
Upon consultation with the teaching staff, a two way ball delivery system that could also hit the Fact Checker seemed impractical. Further, we knew historically that 2 independently driven wheels + 1 or 2 roller ball casters was an effective drive system. Therefore, In the prototyping stage we were primarily concerned with the behavior of the shooter and the ball loading system. We constructed a crude barrel and paddle to load the balls, and borrowed a previous ME210 project to investigate the flywheel shooter method. From this prototyping stage we concluded that using a single flywheel would be sufficient and that the paddle was not reliable enough to load the balls, especially when there was a stack of balls in the tube. So, a large rotating plate would be used to carry and deliver balls so there is no contact between them.
We also prototyped our line sensing circuits (not shown) and tested them on tape attached to white cardboard. While the readings were able to successfully distinguish the tape, being able to line 2 to 3 in a row along a strip of tape led to a larger variation in angle than we liked. Further, the logic for orienting inside the state space using line sensors only proved challenging. We then prototyped two ultrasonic sensors that sat side by side. Originally intended for the Safe Space orientation, we realized that they could be used to run the entire navigation system
Design iterations
Major high-level design changes we made to our initial design, listed in chronological order:
- Removal of two-way shooting (Fact Checker too difficult to activate using Alternative Facts)
- Switch to rotating hopper to load Alternative Facts (Paddle not reliable to handle stack of balls piling down)
- Switch to ultrasonic sensors only instead of line sensors for navigation (Simpler code; better consistency in angle)
- Decision to pursue movable barrel for turret capabilities (Unreliable small rotations from drive motors; is awesome)
Final design
The final design of Gwen consisted of the following features:
- Chassis: Circular body, with threaded rods and nuts to allow construction of additional levels
- Drive: Two independently driven wheels with pillowblocks and rigid couplers; two roller ball casters.
- Navigation: Two ultrasonic sensors, mounted on the same side, co-planar
- Shooting: Single flywheel; PVC barrel, supported by ball caster on barrel end and swivel block on loading end
- Aiming: Servo with two planar linkages, attached to barrel end
- Loading: Rotating hopper that holds 12 balls in separate locations; hole in base plate allows balls to drop through