Abstract
Since 2014, the Formula 1 race car has been equipped with a hybrid electric powertrain combining an electrically turbocharged internal combustion engine with an electric motor/generator unit in a parallel architecture. The energy management system that controls the power unithas a strong influence on the achievable trade-off between lap time andenergy consumption. In our previous research, we developed a robust control algorithm that closely tracksthe lap time optimal operating strategy and adequately reacts to disturbances during the race. Of course, the performance of this controller relies on the accuracy of the models used for optimization. During a Formula 1 race, the vehicle dynamics and the powertrain efficiency may change significantly due to several effects, such as decreasing weight due to the consumption of fuel, variable tire grip due to tire wear and tire changes, as well as reduced engine friction due to heating of lubricants. In order to minimize model mismatch, we present a recursive least squares algorithm to estimate the evolution of the corresponding model parameters and combine it with the previously developed feedback controller. We validate this methodology on a third party high-fidelity nonlinear simulator. The simulation results show that our proposed approach yields close-to-optimal performance in terms of lap time and energy consumption.
Keywords: Energy management, Formula 1, adaptation algorithms, optimal control, recursive least squares