Reliability Tool Overview

​

Channel Compatibility (PCAN + slcan + Damiao Serial Bridge)

• Linux SocketCAN uses interface names directly: can0, can1, slcan0.
• For USB-serial CAN adapters, bring up slcan0 first: sudo slcand -o -c -s8 /dev/ttyUSB0 slcan0 && sudo ip link set slcan0 up.
• Damiao-only serial bridge transport is also available in CLI (--transport dm-serial --serial-port /dev/ttyACM0 --serial-baud 921600).
• Full Damiao serial-bridge interface list and command patterns are documented in motor_cli/README.md (section 3.6 in motor_cli/README.zh-CN.md).
• On Linux SocketCAN, do not append bitrate in --channel (for example can0@1000000 is invalid).
• On Windows (PCAN backend), can0/can1 map to PCAN_USBBUS1/2; optional @bitrate suffix is supported.

• endurance (long-run control/read loop)
• error/timeout behavior (automated unit tests + manual HIL checks)
• disconnect/recovery (manual HIL steps)
• cross-platform scan consistency (Linux vs Windows)

​

1) Endurance

python tools/reliability/reliability_runner.py endurance \
  --command "cargo run -p motor_cli --release -- --vendor damiao --channel can0@1000000 --model 4340P --motor-id 0x01 --feedback-id 0x11 --mode pos-vel --pos 3.1416 --vlim 2.0 --loop 1 --dt-ms 20" \
  --duration-sec 1800 \
  --interval-sec 0.5 \
  --report tools/reliability/reports/windows_endurance_4340p.json

python tools/reliability/reliability_runner.py endurance \
  --command "cargo run -p motor_cli --release -- --vendor damiao --channel slcan0 --model 4340P --motor-id 0x01 --feedback-id 0x11 --mode pos-vel --pos 3.1416 --vlim 2.0 --loop 1 --dt-ms 20" \
  --duration-sec 1800 \
  --interval-sec 0.5 \
  --report tools/reliability/reports/linux_endurance_4340p.json

python tools/reliability/reliability_runner.py endurance \
  --template tools/reliability/templates/linux_slcan_endurance_4340p.json

• tools/reliability/templates/linux_slcan_endurance_4340p.json
• tools/reliability/templates/linux_slcan_endurance_rs00_vel.json

• fail == 0
• success_rate == 1.0

• thresholds.max_fail
• thresholds.min_success_rate

​

2) Error/Timeout Injection

• bus read error path (CoreController)
• Damiao register timeout
• RobStride parameter timeout

​

3) Disconnect / Recovery (HIL)

1. Start a short control loop (pos-vel or vel).
2. Unplug PCAN-USB (or down Linux CAN interface).
3. Verify command pauses, then auto-reconnect/retry is attempted.
4. Re-plug / recover bus.
5. Verify loop resumes without restarting the process, or re-run scan/control to confirm recovery.

​

4) Cross-Platform Consistency (Scan)

python tools/reliability/reliability_runner.py compare-scan \
  --left-log tools/reliability/reports/linux_scan.log \
  --right-log tools/reliability/reports/windows_scan.log \
  --vendors damiao,robstride \
  --allow-hit-delta 1 \
  --id-mode intersect-nonempty

• per-vendor hits with optional tolerance (--allow-hit-delta)
• vendor subset filtering (--vendors)
• configurable id matching mode (--id-mode):
  • exact
  • left-subset
  • right-subset
  • intersect-nonempty