The teams use their departments’ strong points to choose their focus for the four-year project, whether that’s aerodynamics or electronics. The UW team, for example, chose to work on developing a predictive driving hybrid control system that uses machine learning to find more fuel efficient routes. The system learns which trips the driver is most likely to take, say to work every morning, then looks for possible downhills or slowdown events. The car uses these opportunities to use regenerative braking to capture that energy rather than the regular brakes. In simulations with the system this year, they’ve gotten 5-15% better fuel economy.