On Robot Testing

[gerth2]

Per discussions, as of Feb 12, we have the following buckets of work to do. Each bucket must be completed before the next.

Bucket 1 - Component Functionality

Intake

• [x] Position sensor mount, debug, & tune (get to the point where we can accurately read intake position)
• [x] Closed Loop PID position control tune
• [x] Preset Position tune

Arm

• [X] SmartMotion basic functionality checkout
• [x] Limit switch safety operation check in closed & open-loop modes.
• [x] Gravity Compensation Factor with Load Attached
• [x] Velocity PID Tune with Load attached
• [x] SmartMotion velocity/accel tune with Load attached
• [x] Preset Position Configuration Tune - first-pass
• [ ] low/mid/high/intake position tune for Hatch
• [ ] low/mid/high/intake position tune for Cargo
• [x] "Danger Zone" minimum position for arm motion with intake retracted tune
• [x] Initial angle configuration

Gripper

• [x] Basic functionality checkout (confirm solenoid directions & logic assumptions)
• [ ] Position change time profiling

Bucket 2 - Driver Assist Automation - Basic Checkout

• [x] CargoIntake to CargoCarry
• [x] CargoCarry to CargoIntake
• [x] Hatch to CargoIntake
• [x] CargoIntake to Hatch
• [x] CargoCarry to Hatch
• [x] "Chaos Monkey" Safety checkout - ensure rapid commands to change state do not cause robot to enter an unsafe or destructive state.

DRIVER PRACTICE MAY START AT THIS POINT

Bucket 3 - Driver Assist Automation - Optimization

• [ ] For all transitions, do operations in parallel where possible
• [ ] For all operations, observe "dead time" where the robot is not moving. Trim out as much as possible.

Bucket 4 - Autonomy

• [ ] Attach and confirm operation of front and back ultrasonic sensors
• [ ] Complete drivetrain tuning to follow a complex path accurately
• [ ] Test AutoSeqPathPlan can calculate and align to a specific point & angle on the ground (see drivetrain test vector inputs)
• [ ] Confirm operation of distance-to-target-estimation algorithm. Add filtering or priority changes as needed.
• [ ] Confirm JeVois is returning target position and rotation once mounted on the robot.
• [ ] Confirm operation of autoSequencer state machine in "happy-path"
• [ ] Confirm operation of autoSequencer for camera- or path-planner induced errors.
• [ ] Confirm operation of the final path-alignment using angle from JeVois camera
• [ ] Confirm operation of the final path-alignment using line sensors.
• [ ] Tune timing of gripper release and backup
• [ ] Confirm robot actually stops when auto align command is released at any point.
• [ ] Robustness checkout for alignment - explore approaching many targets from many angles, ensure alignment is correct in each case.

[gerth2]

Notably absent from this list is climber. This should be worked in parallel with other testing. The software portion is comparatively trivial at this point.

[gerth2]

shmeh good enough