MONDAY, 14 JULY 1997, THE ATACAMA DESERT
(Llano de la Paciencia)

The Atacama Desert is known as the driest desert in the world. Not to argue with the experts who have declared it thus, but it rained today. OK, a more accurate description would be that it "misted" today, but I definitely felt two drops. The folks observing the panospheric imagery initially didn't believe me when I told them it was raining, but a few minutes later they were asking to have the glass surrouning the panospheric camera wiped off; about 20 drops had hit it. The final touch was something I never expected to see here: a rainbow appeared, positioned just so that the pot'o'Nomad could be found near its base.

Rainbows aside, it is not easy working in the field with Nomad. This afternoon we had scheduled some local tests of the navigation and skyline position systems at noon, so some of us drove out to Nomad to be ready when the time came. Unfortunately the folks at the Science Center spotted us in Nomad's panospheric camera and immediately set after us. We tried backing out of their way, but they were not to be stopped. We had to back our truck over 50 meters away before they lost us in the rocks. Actually, I was already out of the truck with a notebook computer when it started chasing me down, so I had to lie down in the dirt behind the notebook over 20 meters away to stay hidden [helpful hint for Martians and Selenites]. Hans Thomas' comment on 21 June about living in a Terminator movie, hiding in the ruins from the robots, seems especially appropriate again today.

The tests themselves yielded mixed results. The navigation tests were cut short by the unexpected resetting of one of the real time computers. Our real time expert had worked with Nomad all last night, and had just woken up and started checking things out when the system went down and would not be brought back up for over 30 minutes. Funny, how everything seems to work well until the experts arrive. In any case, things were restored just in time for the Skyline navigation tests, which seem to have gone well. The tests were run with the Skyline components mounted on Nomad, using Nomad's power and sensors to determine ground truth location. Two panoramas were taken, and later processing of one of them yielded Nomad's location to within 270 meters, or within 3 pixels on the map. This is a very nice result, especially because Nomad was located in an area that is surrounded by "small" peaks (less than 50 meters) which do not even register in the Digitial Elevation Map used by the method. In spite of those small local peaks occluding the larger, farther ones, the method was still able to pinpoint its location to within a few pixels.

Finally, Nomad has been running continuously from 0700 hours on Sunday through 0400 hours tonight (Monday evening, Tuesday morning). Things went especially well on Sunday, with no real time resets for 15 hours and very quick refueling "pit stops" of less than five minutes. And today Nomad drove itself for 1017 meters autonomously (still with a little help at steering around the obstacles it detected; but automatic point turns are coming soon).


Last Modified on: Tue Jul 15, 1997