We began the day by recalibrating Nomad's stereo cameras in order to obtain settings that work well over the entire range of light conditions (bright mid-day to low-contrast late-day). Nomad's generator, which we plan to eliminate in favor of solar and wind power, needed treatment to remove water/ice that were interrupting uniform power output and causing computers and communications equipment to reboot. Our goal was to prepare everything to run continuously for an entire day.
We chose a route of eight latitude/longitude points that circumnavigated Lake Mascoma and commanded Nomad to execute the route twice, each circuit being about 7 km. The lake surface is 45 cm of solid ice with 15 cm of snow overlaid. There were no physical barriers to Nomad's mobility, just as we expect in Antarctica, but there are a fair number of ruts/tracks that Nomad's Navigator might choose to avoid.
After commanding the path to Nomad we had only sit back and watch for 8 hours. Nomad used its onboard compass and wheel odometry to dead reckon from goal to goal. Upon reaching each goal, the robot would stop simulating the time before and after initiation of science activities. The Navigator used stereo images to model the terrain in front of it to choose an obstacle free path. At dusk when light levels drop too low for vision, Nomad's laser continues to operate and cause the vehicle to backup and recover its path when an obstacle trips the "virtual bumper"--the loss of sufficient light occurred --after about 6.5 hours when Nomad had travelled 11 km.
After just under 8 hours (traveling 0.5m/s) Nomad returned to its starting point, for the second time, having traveled 14078m. It is important to note that the snowy terrain is mostly featureless so the robot was primarily looking for unexpected items in its path. It did this all day long with no need for assistance, demonstrating a first instance of day long autonomy in a snow/ice environment.
Assess autonomous navigation capability on snow/ice. Although we had planned for ice, we arrived at Lake Mascoma after a night of heavy snow. Nomad however had no difficulty with 6 inches of snow onto of the ice.
Initially we experienced communication drops out and controller resets but traced those back to power problems. We found Nomad's internal generator was sagging to 80V and this we due to icing in the carborator. We built a left-handed smoke shifter to redirect the generator exhaust onto the carburator to keep it warm and it has been running smoothly at 120V since.
For the bright snow we've closed camera apertures and recalibrated the stereo vision system. There is enough texture on the snow and ice that Nomad is producing excellent terrain maps, although they are mostly flat. Nomad also carries a laser scanner to act as a "virtual bumper". We have verified improved the filtering on the laser so that it is not confused by falling snow. Other refinements to the navigation system have improved tracking of heading by more finely resolving steering arcs.
Integrate wind turbine. We have conducted a simulation study to solar and wind energy in Antarctica to determine whether our intended test area near Carapace Nunatak will be able to support a continuous solar/wind rover mission. We've installed a wind turbine on Nomad to assess stability and measure output. An anemometer gives us continuous wind speed which we can use to correlate turbine power to our model.
This research is supported by NASA.