Race car performance evaluation by a neuro-fuzzy inference system

In this paper, a tool supporting racing car design and set-up is proposed. A neuro-fuzzy inference system is used to model performance using set-up parameters. This approach has been suggested by the need of adopting a linguistic description of the system itself. The network has been developed by performance simulations of a race car in known conditions. A variational analysis of the output surfaces can provide a simply way to achieve the desired car performances