A Novel Strategy Model Approach for Racing Solar Cars in the Bridgestone World Solar Challenge #SWI2023
Top Dutch Solar Racing is a student team from Groningen that aims to develop and test technologies in the field of sustainable car mobility. We do this by building a solar energy-powered electric vehicle. The goal is to contribute to the industry by using and validating new technologies to improve carbon emission-reduced mobility. As a proof of concept, the team competes in the Bridgestone World Solar Challenge, a race in which teams have to drive 3000 km through the Australian outback using the sun’s energy.
Driving a good race is about a few things: the quality of the car and the quality of the strategy model are arguably the most important. The strategy model tries to work out a plan to cover the race distance in a minimal amount of time. At its core, the strategy model suggests driving speed and corresponding power output for each route segment, based on a Newtonian physics model. The strategy model on the lowest level considers the amount of energy in and the amount of power going out of the system.
The Mathematical Challenge
Top Dutch Solar Racing would like to build a new strategy model from the ground up, using the principles of a (non-linear) model predictive control (MPC) system. Therefore, the first challenge is finding a reasonable running cost function based on the physical model. Furthermore, deciding on an efficient search algorithm is paramount due to the fast-paced nature of the disturbances in the race. Lastly, choosing an appropriate method of discretization should be valued to ensure high robustness and accuracy of the model.
Solving these problems will allow Top Dutch Solar Racing to make the best use of its computational power, ensuring the strategy model can suggest driving speeds with high resolution. Hence, the team has a good chance of driving their best possible race.