Day 3 started with compass problems, the compass is necessary for a boat to know which way the boat is orientated. The compass used is a 3-axis magnetometer and is very sensitive to small magnetic fluctuations. Being inside a steal framed building or having the boat on a steal table can significantly impact the compass's direction. However, even outside and away from all potential magnetic disturbances the compass was giving strange readings. In the end we discovered that there was a bug in the program and it nothing to do with the compass itself. A few hours had been spent trying to find the problem and it took two minutes to solve it.
The good thing about the boat that was leant to us by Ensta (the French team) was that it came with a wind vane at the top of the mast. With some help by Mark from The Neils (the Welsh team) we were able to integrate the wind vane into the computer. This is a massive improvement over the previous wind vane which was kept on shore. The wind vane is important because it tells the boat the wind direction. While an anemometer will give the boat the wind speed, however the wind speed is not available with the new setup, this is not actually a big concern, since wind direction is the most important. Overnight we were able to complete the integration of the onboard wind vane with the artificial intelligence.
Day 3 is the Station Keeping contest. The idea behind station keeping is to keep the sailboat positioned in a given area for 5 minutes. The conditions on the day were not the best, the current was high and the wind was strong. However the boat did well, we came 3rd. Although looking at the logs afterwards the GPS signal was lost half way through, so the boat went off in a straight line. See the plot of our station keeping attempt below. The reason it went away in a straight line is because when it doesn't have GPS it thinks the speed is zero and just keeps going.