Our robot can successfully aim and hit targets at 5, 10, and 15 yards. Table 1 shows our accuracy across 3 trials at each distance, with 3 distinct targets, with no adjustments.
Table 1: Accuracy Results
| Distance (Yds) | Trial 1 | Trial 2 | Trial 3 |
| 5 | 2/3 | 3/3 | 3/3 |
| 10 | 3/3 | 2/3 | 3/3 |
| 15 | 1/3 | 3/3 | 2/3 |
Table 2 depicts the same results with the opportunity to make 1 adjustment (either up, down, left, or right).
Table 2: Accuracy Results
| Distance (Yds) | Trial 1 | Trial 2 | Trial 3 |
| 5 | 3/3 | 3/3 | 3/3 |
| 10 | 3/3 | 3/3 | 3/3 |
| 15 | 2/3 | 3/3 | 2/3 |
Why isn’t it perfect?
The accuracy results were not perfect, which wasn’t completely unexpected. We do believe this results from two main factors:
1 – The camera was not perfectly level
2 – The pitching machine was inherently inaccurate
In terms of the camera, we noticed two things after taking some initial data, it was not perfectly level. It was not only pitched up slightly, but also very slightly crooked as well (with the left side being higher than the right). This led to the pitch and yaw angles being slightly off when aiming at some squares, in our case typically shooting a bit too low at targets more to the right than to the left. In our demo video, you can see this when it aims at target 2. Upon shifting our camera to make it more level, we noticed an improvement in accuracy.
The pitching machine, being a single flywheel design, was also not the most consistent piece of equipment. We see from Table 1, when shooting at a 1ft by 1ft square, it did miss randomly when aimed at the square with no adjustments. We do believe that no matter how lined up the launcher is with the center point, it will still miss sometimes due to inconsistent loading. The speed in which you put the ball in, and spin it has, and even how it is shaped plays a huge role in accuracy of the single flywheel launcher.
We did an accuracy test with the single flywheel launcher where we manually centered it and shot it at one target multiple times. The results shown in Table 3 depicts the inherent inaccuracies of the launcher. This is why we build in the ability to slightly adjust after each shot.
Table 3: Accuracy Experiment Results
| Distance | 1ft Square | 2ft Square |
| 5 yds | 18/20 | 20/20 |
| 10 yds | 11/20 | 20/20 |
| 15 yds | 7/20 | 19/20 |
| 20 yds | 4/20 | 11/20 |