Testing Report: July 12th 2001 < Back
 
Devon Island, Nunavut, Canada

Testing started: 9:30 am
Testing finished: 12:30 pm

Agenda:
- Mount pyranometer on robot.
- Perform stereo component testing.
- Perform laser component testing.
- Navigate to waypoints using Hyperion's autonomy system.
- Tune panoramic camera settings.

Progress:
- Tuned shutter speed, white balance and exposure control on panoramic
camera, although perhaps it could be manually focused to get better results.
We may need to re-tune in different lighting conditions.

- Captured 5 sequences of stereo images of robot approaching a 47 x 37 x
18 cm rock. The time was 10:10 am and the weather conditions were partly
cloudy with a bit of sun.

- Pyranometer attached, although PMAD code must still be updated to display pyranometer data.

- Robot drove to several waypoints very well. Experienced probably four or five GPS jumps. Stereo reliably saw Ben. Ben also triggered an E-stop correctly by tripping the laser. At that point he also correctly caused a NO_ARCS fault from the Navigator. Therefore the Health Monitor seemed to be responding to
two common faults properly.

- Fixed a bug in the Mission Plan Tool of the Operator Interface that incorrectly displayed data when the Navigator reached a goal.

- Took laser and stereo data of robot approaching an obstacle that was missed during autonomous navigation. Used these data to tune stereo and laser parameters, hopefully without introducing too many false obstacles.

Problems:
- As mentioned above, neither the stereo nor the laser could detect a 30-40 cm rock in front of the robot's right wheel. This should have flagged an E-stop, but instead Ben hit the robot's E-stop switch. See analysis results, above.

- Saw a few more cases of amp 1 disabling, though not as many as yesterday. Also had trouble backing up onto a rock once.

To Do:
- Test stereo and laser again in the rock field with the new settings.
- Download new PMAD data to unit on robot and test that pyranometer data are valid.
- Install panoramic camera on robot.
- Continue battery discharge data analysis, integrate into Mission Planner.



   
  ©2001 Carnegie Mellon University - Robotics Institute