Genetic Algorithms Applied to Determine Optimal Trajectories of High Speed Ground Vehicles in Predefined Tracks. 2008-36-0131
This research work details an optimization method to identify the path of a pre-defined track through which a car with specific acceleration limits completes the circuit in its lowest time. The strategy adopted here uses the evolutionary approach of a Genetic Algorithm's based optimization method to determine the accelerations suffered by the vehicle through the circuit, from pre-defined profiles. Considering an Oriented Particle vehicle model, those acceleration profiles can be integrated to find a specific trajectory path.
The results are quite relevant compared with other methods used, such as the Optimal Control and the conventional Least Squares optimization. Receiving as the initial estimation the lane centerline of the track, the first Genetic Algorithm tests managed to reduce the lap time in about 10%.