We spent the night before the scrimmage squishing a couple of bugs:
- The optical distance sensor would return near-0 values when the robot approached the border, causing our autonomous to fail. This was determined to be a result of the sensor being slightly tilted and the mat having a slight bump near the edge.
- We had issues with all sensors mounted on the core device interface module failing to read. After running our test programs and switching cables/modules around, we found a broken USB cable connecting the sensor module to the power distribution module. After replacing this cable, the module worked properly during the scrimmage.
- The gyro sensor would lead to crashes and sometimes get stuck while trying to calibrate. This is a known problem in the FTC SDK and was fixed in the 2.61 release, which we merged into our own repository.
Scrimmage notes and future work:
- Autonomous is mostly consistent and during some of our matches we narrowly missed the beacon/center ball shooting. This should be easily fixed with another day of tuning and some mechanism adjustments.
- We need to update the code for the new Actobotics chassis, which was recently completed, and write code for the new sweeper, cap ball lift, etc.
- We need to add encoder limits for the cap ball linear lift to prevent the lift string from snapping when it got over-extended during practice.
- The test programs/telemetry reading works well, but additional telemetry before the autonomous starts to monitor sensors would be useful in making sure that all of the sensors are working properly.
- We should write a full set of autonomous programs, including delay choices and only shooting center balls.
- Fix the slow mode toggle in Teleop using the technique here, or better yet rewrite the entire joystick algorithm (see here).
- Continue to develop Vuforia and other computer vision algorithms.
12/17/16 Software