Virtual coach for super sports cars: a possible AI application

Chiloiro, Giuseppe Marco (2021) Virtual coach for super sports cars: a possible AI application. [Laurea magistrale], Università di Bologna, Corso di Studio in Automation engineering / ingegneria dell’automazione [LM-DM270], Documento full-text non disponibile
Il full-text non è disponibile per scelta dell'autore. (Contatta l'autore)


The dream of the human being has always been to reach the perfection. This is true for life in general, but also for sports competitions including motorsport. In car racing, the optimum is represented by the perfect lap, performed by the driver in the minimum possible time, following the ideal race line with the appropriate velocity. However, due to the intrinsic nature of human being this is unfeasible and the only way to get as close as possible to the perfection is to rely on technology and artificial intelligence. Focusing on motorsport, the actual knowledge of vehicle dynamics and the development of optimization techniques can help in reaching great improvements in terms of race performance. This is exactly the objective of this thesis in which the implementation of a virtual coach for super sports cars has been carried out, in order to provide useful suggestions to the driver during and after the performance. In particular, the algorithm on which the virtual coach is based, computes the ideal race line and the optimal velocity in every point, giving also an accurate estimation of the best reachable lap time. The virtual coach is able to adapt to each super sports car and circuit, since the vehicle and racetrack models are generated starting from the physical parameters introduced by the user as inputs to the algorithm. The optimization of the race line is performed by solving the minimum lap time problem through the optimal control theory and the optimal velocity is computed for each point of the ideal trajectory. After the first lap, the algorithm associates the driver to a category level and shows which are the sectors of the circuit where a higher improvement is possible, adapting the targets in terms of lap time and velocity to the driver’s capabilities and allowing a step by step improving of the performances. The algorithm has been tested on a large variety of vehicles and circuits, providing accurate results in terms of ideal race lines and velocities.

Tipologia del documento
Tesi di laurea (Laurea magistrale)
Autore della tesi
Chiloiro, Giuseppe Marco
Relatore della tesi
Correlatore della tesi
Corso di studio
Ordinamento Cds
Parole chiave
Data di discussione della Tesi
10 Marzo 2021

Altri metadati

Gestione del documento: Visualizza il documento