A Feasibility Study on Driver Model Based Lap Time Simulation Using Genetic Algorithms

Lap time simulation has always been a topic of interest in the automotive industry as it summarizes the whole dynamic performance of an automobile in a single value. During the development of road and race cars, to avoid expensive testing and to prove different design solutions, it is useful to simulate the maximum performance of the vehicles. The cars are driven to their limits to exploit their capabilities, where their dynamic behaviour can be highly non-linear. The vehicle models need to replicate these characteristics as precisely as possible. Due to this, the problem of achieving the minimum lap time with a certain car around a race track is a challenging problem to solve. A method to evaluate the minimum lap time is presented, approaching the optimal solution by coupling a driver model, a simulation environment and genetic algorithms to perform the optimization. The algorithm also offers the possibility to add vehicle parameters to be optimized regarding the lap time. The process can be adapted to different vehicle, track and driver models. Not being limited to a single simulation environment, the high flexibility of this method allows it to work with diverse software. The optimization algorithm uses a newly developed penalty function to improve the convergence and accuracy. Real tests on a race track have been performed to assess the correlations between the measurements and the suggested method, revealing its high accuracy.