Mars or Bust

The Mars Rover Competition is over! Finally! Don’t get me wrong, it was a great experience, and I’m glad I did it, but it was a lot of stress, a lot of time, and a lot of frustration.  Despite having the highest score in the Autonomous Traversal task, we ended up placing 5th at the 2018 University Rover Challenge in Hanksville, Utah, which considering all the problems we had, is quite remarkable. After leaving work a bit early to make it to the van at 5:30, I arrived on Wednesday evening to find the rest of the team furiously working on the rover. They had arrived earlier in the morning and as they were unloading, a bad solder job on the power bridge caused a blowout and fried two USB cameras and the USB hub. Along with swapping in the spares to address this problem, the mechanical team had waited until the last minute to buy new tire treads, so half the team was working on removing the tires and putting the new treads on. We stayed up until 1 a.m. getting everything fixed, then got up around 6 to get the rover ready for the first task: Extreme Retrieval.

We were feeling really good after testing the rover at camp, but when we arrived at the task site, we were dismayed to see that the rover had not been tied down correctly and was seriously damaged. The turret motor for the arm was completely broken and the front wheels were bent. We had 20 minutes to prepare for the task, so we furiously tried to replace the turret, but couldn’t do it fast enough. We ended up tying down the arm to face straight and figured we could just turn the rover itself to get the same effect. Starting 10 minutes late into our task, our spirits soared when the rover picked up a toolbox and started heading towards the dropoff zone. They sank again however when the rover stopped 1 meter from its goal, turned around, went up a little canyon and proceeded to drive up the wall and flip over. We couldn’t believe what had just happened! Apparently, the judges had given them a bad GPS waypoint to the dropoff zone, and they trusted that point more than their camera feed. We performed an intervention by picking up the rover and turning it back on, which cost us 20% of our points. Then the rover continued on, driven by our operators in the Hab, and picked up a water bottle and dropped it off in front of an astronaut (mannequin) right as time ran out. To go on to the second leg, we needed 40 points, which we had, but the intervention dropped us down to 32 according to the judges, so we had to stop right there. We were really disappointed as we got back to camp, and spent most of the day fixing the turret motor, double checking our backups, and preparing for the next day’s tasks: Equipment Servicing and Autonomous Traversal.

We double checked everything before leaving bright and early at 6:15 a.m. for the task site. Equipment servicing was first, and things fell apart quickly about 5 minutes in when the wrist motors died. Someone had forgotten to plug them in correctly so they got tangled up and pulled out. Somehow, we managed to squeeze 26 points out of the task, but it was painful to watch. However, when we started up Autonomous Traversal, we were surprised to have no hiccups. Everything worked just like we had practiced! The tennis balls were placed farther away from the GPS waypoints than we had anticipated, so we changed the code on the fly to expand the search radius to 15 meters instead of 5. Obstacle avoidance worked fantastically and the InertialSense gave us good heading and position data. The ZED recognized tennis balls better than we had ever seen, and the GUI never froze or malfunctioned. We successfully reached 5 tennis balls, and the only reason we didn’t reach the 6th is because it was 30 meters away from the GPS waypoint. The judges all had smiles on their faces because most of the rovers up to that point couldn’t complete a single leg. Our score of 55 remained the highest of the competition, and the nearest competitors got 40. Only 8 of 35 teams even scored at all. We were pretty proud of that accomplishment, and felt even better when we learned that the judges were going to allow us to try the second part of Extreme Deliver and Retrieval that afternoon. Everything was back up in working order, so I even made some time for Spike Ball after lunch. We arrived at the site and set up the rover, but didn’t have time to check all the motors, two of which decided not to boot up correctly. If we had realized this, we likely could have restarted the rover and fixed the problem, but instead we started into our time with only 4 of six motors working. The operators seemed to have trouble maneuvering, but were able to get 10 extra points by picking up a hammer and dropping it off by another astronaut. However, the last 30 points available to us required scaling a steep hill to a third astronaut, and without all of our wheels in working order, it proved to be impossible. The rover made it nearly to the top, but couldn’t mount the last bit and started sliding down the hill, seemingly to its doom. Luckily, Alex Jensen sprinted after it and caught it before it could tip, sliding down with the machine into the canyon below. Luckily neither he nor the rover were hurt!

The next day, we had the science task, and our scientists were definitely stressed. However, the science module seemed to work right when we tested it the morning of the task, and after mounting it to the rover and tying the whole thing down tight, we drove to the site with optimism. However, when we arrived, we saw that mounting the science module before traveling had been a bad idea. It was too heavy for the wheels to take, and the back two were bent terribly. In order to keep the total weight under 50 kg, the science team only uses four wheel instead of the normal six, so we had already taken off the two wheels that had bent during the previous trailer fiasco. The science module needed straight wheels to dig correctly, so we pulled the unharmed front wheels off to replace the unusable back wheels, then put the remaining two, slightly bent wheels, on the front. It seemed to work, and we got ready for the event. Then the no-nonsense science judge kindly informed us that she had decided to cut the allotted time to dig for samples from 20 minutes to 15 minutes. The plan was to start and finish at a specified point, collect pictures of an anomaly the judges had placed in the area, stick a probe into the soil, and dig two samples and seal them in urine sample containers (unused of course). Everything was going peachy until we tried the probe. The soil was so hard that instead of forcing the probe into the dirt, the linear actuators lifted up the rover and bent the probe. Luckily, they were able to collect measurements regardless, and the probe came out of the soil after lots of pulling. Then, when digging for samples, up went the rover again. Luckily the auger was strong enough not to bend, but it took three tries before we finally found a location where the soil was soft enough to dig a sample, and a small one at that. When presenting the results, our science team answered all the questions deftly and looked very professional, labcoats and all. It turned out that we received a 95.65 on the science task! We were thrilled with the result. Unfortunately, it was only enough to land us in fifth place. The competition really was a lot of fun though, however stressed we felt at times. It was a week of robots on Mars for crying out loud!