Endurance Testing

We've been working furiously to finish up the project before our scheduled departure date back to the states. I've got a bit of catching up to do with the blog, since we've been doing a bunch of runs and tests since the weekend. After getting back from the Shear zone, our highest priority was in showing that Yeti could complete a fully autonomous and uninterrupted run without any input from the users. The test in the shear zone was mostly successful in that Yeti operated reliably and completed its fun while collecting the radar data we were looking for, but we had to stop it a few times and I made several mistakes that caused it to stop and restart. We wanted to show that we could complete a course longer than five miles entirely hands off, so I worked to set up a course on the sea ice near to McMurdo.

We have several challenges to staging this test. The first, is that prior to this deployment, we had not had the opportunity to fully test the radio range, and our observed range was MUCH lower than what we were expecting based on the manufacturer's claims. The datasheets said 60 mile range, but that apparently assumes that each antenna is 100' off the ground, which is obviously impractical for us. Instead, we were finding that we could only reliably communicate out to 3/4th of a mile, which is basically worthless when we're operating 5 miles away.

I set up a few tests and was able to get reliable communication at 3 or 4 miles by raising one antenna a lot higher off the ground and sending very short messages. We only had one day to perform the final endurance trial for Yeti, so everything was coming down to the wire. I wanted to make sure that we wouldn't lose radio comms under any circumstances, so I had to MacGyver together a tall transmitter and pole. The final design used the following parts:

9 bamboo poles
1 roll of duct tape
1 sharpie
2 hoseclamps
2 chemical handwarmer packets to keep the batteries warm
1 chemical spill blanket to insulate the batteries
1 piece of rope from the crevasse rescue bag
4 serial cables, one of them homemade




My kludged antenna mounted on the living module.


The final antenna pole was somewhere between 25'-30' tall when mounted on the living module for the traverse. The rest of the traverse team was laughing at this epic kludge, but it turned out to be brutally effective and gave us a significant improvement in range.

The other major constraint was that we didn't have access to a truck since another 50 scientists have shown up and resources are pretty tight. They've started packing 3,4, and even 5 people to a single room in some of the dorms, but luckily Jim and I have escaped that mess by virtue of the fact that we're leaving in a few days. Not having a truck to chase Yeti makes everything a bit harder, since we wouldn't have any ability to stop or fix it if there were any problems on the run. We were confident enough that it would perform well that we weren't too worried about it, but I still wanted to be there in case something happened. The only way I was going to be able to keep up with Yeti was on skis, so I got a pair of skate skis and took them down to the traverse staging area.

I haven't cross country skied in over seven years, and haven't ever tried skate skiing before, so this was going to be a bit of an interesting challenge. Ben Koons, a friend from Dartmouth, skied for New Zealand in the Olympics, so I assumed it wasn't too hard if he had gone that far with it (Hi Ben). I tried double-polling for about 100' and it seemed pretty easy, so we set Yeti up and sent it the waypoints. I called to Jim on the radio to press the play button and we were off.

The first several hundred meters were a breeze, and I had no problem keeping up with Yeti despite my total lack of any skill on nordic skis. After about 500m, I started slowing down appreciably, and noticed that Yeti was starting to get out ahead of me a bit. The full course length was 6.7 miles, so I began to get a little nervous. It got A LOT worse over the next couple miles, and Yeti proved to be a totally unrelenting pace car. Somewhere around mile 2.5 I started to get the hang of skate skiing, at least to the point where I could pour on the heat and sprint ahead of Yeti, then take about five seconds to recover and start slogging along again.

I made a mistake while getting the coordinates and hadn't fully appreciated the slight bend in the road to Pegasus. This meant that Yeti had to cross through a line of bamboo flags twice both out and back along the route, and I was worried that it would get stuck on one. The first time it crossed the line, I was heaving to keep up with it and couldn't get to it in time as it was heading straight for a flag. Yeti managed to snap it in half without any problem, and continue on, perfectly threading its way back across the line and narrowly avoiding another flag on the way back. We might need to install some vision systems in future designs, but for now it seems to work to give Yeti enough power to mow over whatever lies in its path.

Right around mile 2.3, it made a turn towards a waypoint and looked like it was heading home, which gave me a huge sense of relief, but then continued on its way and I realized I was only a third of the way into the course. After what seemed like an eternity, we reached the turnaround and Yeti successfully hit all of the waypoints and started back home. This time I was more determined to make sure Yeti didn't hit anything, so I tried as hard as I could to stay out in front. It managed to avoid everything else along the way, and make it back home.

I was totally wiped out and Jim was laughing at me when we got back, but the test was a total success and the first time that Yeti had finished a long distance course without any input or restarts.



I'm a little too addicted to the soft-serve ice cream here and eat it even outside. I had an interesting experience yesterday where for about half a second I was wondering what was going on when I realized that my ice cream was getting harder and unmelting after I walked outside into the -20 degree air. It kept getting more solid until it became difficult to eat, which was definitely the first time that had ever happened to me!


The sun partially occluded by Observation Hill on the way home from Scott Base to McMurdo.