| Issue Number | Date Found | Date Fixed | Change Made By | Origin | Description | Resolution | Artifact(s) Changed |
|---|---|---|---|---|---|---|---|
| 1 | 2-09-20 | Shasa | ARM Architecture Device | Issues with installing and running packages and pre-built deep learning models on Jetson Xavier | Load models into a docker image for the Xavier | ||
| 2 | 2-18-20 | 3-28-20 | Bryson and Shaun | Unable to send PWM commands from Xavier | Due to issues using NVIDIA's proprietary language to send commands, we have been unable to generate signals from the GPIO pins | Use the Xavier to communicate with a PWM board over I2C, which will then be used to generate control commands on the car. | Another electronic device to fit into our enclosure and wire up |
| 3 | 2-25-20 | 4-30-20 | Wesley | Faulty USB Port on Jetson Xavier | Could not use certain devices on USB port, which lead to data transfer issues when splitting off from a USB-C port | Ship back to NVIDIA for replacement; under warranty | Have to reinstall and reconfigure device |
| 4 | 2-26-20 | 3-31-20 | Bryson | Unstable lateral vehicle controls | During path tracking simulation, the vehicle was unable to track the path in the lateral direction without going unstable | Change control method from linearized state-space method to Stanley control method for now. | Outputs from controller have changed |
| 5 | 3-11-20 | 3-20-20 | Team | COVID 19 Pandemic | Lost guarantteed access to the lab space and working together in-person | - Zoom meeting calls - Move equipment out of the lab - Alter SVD demonstration plans | Basically everything was changed in some way |
| 6 | 4-15-20 | 4-20-20 | Bryson | Bug in tracking controller error calculation | A bug in the calculation of the lateral error was causing the controller to be driven unstable in certain circumstances. | Changes error calculation code | Conclusions drawn for performance of controller before this were all incorrect |
| 7 | 4-13-20 | 4-15-20 | Shaun and Wesley | Incorrect TVS diode used | PCB traces blown | Scrape off traces and solder jumper | Rewired board to make it functional |
| 8 | 4-22-20 | 4-25-20 | Evan | Sub-optimal distance metric | Lateral oscillation in the planned trajectory in certain scenarios | Improved distance metric that incentivizes facing down the road. | Which distance metric the planning system used. |
| 9 | 4-20-20 | 11-10-20 | Shaun and Evan | Uncalibrated / fully complete sensor integration, potentially | Onboard system is unable to track a given trajectory | Fixed issues with the jacobians used for some sensors and removed contributions from IMU that were degrading performance. | Likely localization and sensing systems. |
| 10 | 8-30-20 | 10-15-20 | Shaun and Evan | Encoder magnetometers | Magnetometers keep erroring at a high frequency and output junk data afterwards. | Better wiring and build in a recovery mode in the Arduino code to power cycle the magnetometer and reestablish the data connection. | cable harness to magnetometers redone |
| 11 | 9-10-20 | 9-20-20 | Bryson | GUI freezing | GUI not able to function with multithreading, so certain processes can't be spun up by the GUI | Take those functions out and generate a separate bash script to run the ROS processes on the robot | More inconvenience during testing |
| 12 | 10-01-20 | 10-10-20 | Shaun and Evan | IMU yaw values inconsistent | Yaw value reported by the IMU would change with position and not return to where it was previously. | Turned off the compass contribution to the yaw value. | IMU |
| 13 | 10-01-20 | 10-10-20 | Evan and Bryson | Controller going unstable | When attempting to track simple slow speed paths, the vehicle consistently goes unstable | Fixed bug in code where positive feedback look was occurring | No negative consequences |
| 14 | 10-05-2020 | 10-15-2020 | Shaun | FCHarDNet frame rate too slow | The frame rate for outputting segmented images did not meet our necessary speeds. | Multithread the image loading and segmentation portions of the code | Terrain Comprehension |
| 15 | 11-05-20 | 11-10-20 | Bryson and Evan | Planner would failing | One set of motion primitives could either plan a short horizon or a long horizon, not both. | Have two different sets of motion primitives. | More work to switch between online and offline planning modes. |
| 16 | 11-12-2020 | 11-16-2020 | Team | Crash | The car crashed at full speed into the guardrail. | Buy a new GPS antenna. | The platform |
| 17 | 11-23-2020 | 11-23-2020 | Team | GPS Broke | GPS board would no longer connect to our computer or to the other GPS board. | Get a new GPS board | RTK GPS system |
