The Extra-Vehicular Activity Robotic Assistant (ERA) team, consisting of Jeffrey Graham, Robert Hirsh, Nathan Howard, and Kimberly Shillcutt completed the two week field tests in cooperation with the Mobile Agents/Brahms team with significant accomplishments.

Our simplest measure of integration success, based upon how well we provided the functionality as specified in the requirements, was 88% (8/9) during the final week of testing. In essence, the ERA team was required to provide:

1. Push GPS: the ability for the Brahms system to add Global Positionining System (GPS) coordinates for named targets/waypoints into the ERA robot system.
   
   
2. Get GPS: the ability for Brahms to ask ERA for the GPS location for any known target/waypoint.
   
   
3. Stop Move: the ability for Brahms to stop ERA's tracking movement.
   
   
4. Move To: the ability for Brahms to command ERA to move autonomously to a known target/waypoint.
   
   
5. Take Picture: the ability for Brahms to command ERA to take a picture of a given target, and store the image permanently in its database with salient metadata including timestamp and description or name.
   
   
6. Get Picture: the ability for Brahms to re rieve an image from the ERA database given the metadata.
   
   
7. Tracking: the ability for the ERA software to dynamically track the robot, waypoints, and astronauts in a Graphical User Interface.
   
   
8. Video: provide video feed, image archival, and camera pan-tilt control onboard the robot to habitat scientists to enhance their ability to assist the remote "away" team.
   
   
9. Situational Awareness: provide distance and bearing information on request between robot, waypoints, and astronauts.

The only failure of the ERA was the ability for the astronaut to command the robot to move autonomously to a named waypoint. Although initiating this action through the Brahms-ERA interface failed, this functionality in fact existed and repeatedly succeeded during the tests when so initiated via other means.

Additionally, the ERA provided alternative communications capability via speech synthesis of text messages from the MDRS habitat through the robot for the astronauts. The ERA robot carried sample bags for the astro-geologists as they collected, bagged and tagged interesting samples. In several scenarios, the robot acted as a data communications relay between the astronaut suit computers and the wireless backbone connecting all computers together.

The ERA team demonstrated lessons learned from previous field exercises, with improved battery management, logistics, and effective contingency mitigation. System software was more mature and more capable. The ERA experienced fewer problems related to wiring, power, and sheilding from radio frequency interference with respect to the GPS, spacesuit backpack, and inertial measurement sensors.

From this field test, new lessons have been learned:

1. Every single point failure must have a spare. Boudreaux was immobile 2 days because a chain sprocket broke and the nearest spare was a 5 hour drive away.
   
   
2. More autonomous behavior is useful when communications dropouts are frequent. Boudreaux was very robust during the communications dropouts.
   
   
3. All connectors should be secured with screws or clamps to prevent disconnects. Several disconnects cause minor troubleshooting delays.
   
   
4. All connectors should be labeled on both sides, and where the possibility exists for damage, unique connectors should be used to ensure correct connections are made. Just days before the field tests, a wrong connection resulted in burnt electronics that took a week of significant effort to recover from.

ERA had several team objectives as well:

1. Multi-waypoint tracking, with features such as pause/resume/abort, using global or relative poses:
   This was quite successful and substantially improves the autonomous capability and team-player functionality for the ERA robot.
   
   
2. New mobility base testing:
   The second generation ERA was successfully tested during this field exercise. Intended to eventually replace the current ERA, this vehicle is more stable and is able to negotiate more difficult terrain. Testing was conducted using a person to simulate the payload weight and intelligent control system.
   
   
3. Astronaut tracking using the BVS, pushing pose to the poseServer, providing image capture on demand:
   A hard disk crash on the BVS system prevented extensive testing in this field exercies. The ERA team was able to resurrect the hardware and software, and for the last several days of the field exercise, the tracking worked well in a benchtop/lab environment.
   
   
4. Astronaut tracking using the spacesuit infopak with high precision GPS:
   This capabality has been testing in the past through the same software code paths mentioned in #1 above.
   
   
5. Obstacle detection & avoidance using a laser rangefinder:
   This new functionality was subjected to initial testing in this ultimate environment; extensive data was collected, several bug fixes were made, and the robot recognizes and attempted corrections on its way to GPS locations. Tweaking parameters and fine tuning must still be done upon analysis of the data.
   
   
6. Path planning using a robot-generated "goodness" map

Experiments were successfully performed for most of these objectives, while valuable field test data was gathered for the remaining objectives.

To summarize, the ERA team provided most of the requested functionality and several extra functionalities that proved useful for the astro-geologists and scientists working in the Mars Habitat. The robot performed well in the challenging environment including the rough terrain and difficult wireless communications.