| Issue Number | Date Found | Date Fixed | Change Made By | Origin | Description | Resolution | Artifact(s) Changed |
|---|---|---|---|---|---|---|---|
| 1. | Sep 12 | Sep 26 | Boshen | Large difference between current and desired angle | "Short" error detected on motor controller GUI due to large current | Added interpolation function before sending motor commands | Motor controller code |
| 2. | Mar 12 | mar 28 | All Team members | Interfacing motor controllers | It proved to be very difficult to interface the motor controllers using etherCAT | Switched to Roboteq motor controllers that communicates with ROS | Motor Controller |
| 3. | Nov 8 | Nov 9 | All Team members | Jetson | issues with GUI, internet, and unreliable results | switch to off-board (laptop) till pipeline ends | Jetson |
| 4. | Sep 12 | Sep 26 | Rupika | Swing leg prediction algorithm | exec time > 100 ms | Changed model to logistic regression | python code |
| 5. | Oct 15 | Oct 18 | Aubrey | Electrical Wiring | 48V connections to power distribution block were loose causing motor under voltage when they disconnected | Tightened Quick Connect connectors that connected the wires to the power distribution block so that they did not come loose with movement of the exoskeleton | Tightened Electrical Wiring Connections |
| 6. | Aug 30 | Sep 11 | Aubrey | Motor Controller | Discovered Motor controllers cannot handle large amounts of reverse current generated by the motors when the controllers are powered from a power supply | Added a power dissipation circuit that would dump the excess current to ground when an over voltage was detected by the motor controllers | Added Electrical Power Dissipation Circuit |
| 7. | Oct 20 | Oct 24 | Aubrey | Motor Controller Wiring | Many of the motor wires were disconnecting from the motor controller | Added strain relief to the motor wires and several other wires as well. Also tinned the ends of the wires so they fit more securely into the motor controller terminals. | Improved Electrical Wiring Robustness |
| 8. | Nov 5 | Nov 15 | Angela, Sachin | Secureness of exoskeleton when user has it on | Exoskeleton was "drooping downward" from its weight after a period of wear. | Shoulder straps were ordered and installed on the main frame of the exoskeleton to ensure vertical component of the weight is supported sufficiently when the user is wearing the system. | Exoskeleton frame |
| 9. | Nov 15 | Nov 17 | Boshen | Motor Controller Current Limit | High current limit on the motor controllers meant that the user for on person testing could not over power the motors. This was a safety concern. | Lowered the current limit of the motors to 5Amps so that the user could over power the motors as a safety precaution for on person testing | Motor Controller Setup |
| 10. | Nov 4 | Nov 5 | Sachin | Fixed precompute did not fully handle the situation | When interpolating the normalized winter trajectory inside matlab before porting to C++, it is not possible to quickly change the number of datapoints to interpolate so that testing can be done smoothly. | Implemented an interpolation function inside the C++ code instead of the matlab code | Trajectory planner function |
| 11. | Nov 9 | Nov 11 | Angela | Double swing cases | During double swing, currently both legs become free-spin, however we only want the second swing leg to be in free spin while the leg that went into swing first to hold its landing angle position until touchdown | Added code to determine and track which leg went into swing first, then wrote the output cases. | Capture Point Trajectory Methods |
| 12. | Nov 11 | Nov 11 | Angela, Sachin | Vector buffer was not being cleared | The vector of the winter trajectory was being pushed-back each time the planner was called but was not being cleared once a swing was completed | Added a vector clear function in the initial portion of the trajectory planner | Trajectory planner function |
| 13. | Nov 20 | Nov 20 | Sachin | Phase shift | Phase shifting the trajectory also caused it to scale along the x-axis due to a mismatch in the number of datapoints when phase shifting | Modified the number of datapoints to be a function of the extent to which the trajectory was phase shifted | Trajectory planner function |
| 14. | Nov 24 | Nov 24 | Sachin | Interpolation size | The number of waypoints being used were too large | Tuned the number of waypoints by testing the motor response | Trajectory planner function |
| 15. | Nov 25 | Nov 30 | Angela | Adduction/Abduction control | Output trajectory not correct in terms of magnitude and directionality | New algorithm implemented to fix target magnitude calculations and ensure correct directionality | Capture Point Control Method for Adduction/Abduction |