Braking torque estimation through machine learning algorithms

Abstract. MotoGP class motorcycles rely on carbon braking system to cope with their incredible acceleration capability and high speed. Hence, assessing the torque generated by front discs is a key improve vehicle performance. As direct measurement of not allowed during races, its value may be estimated through physical model, using as inputs brake fluid pressure (monitored board), geometry friction coefficient (μ). However, results obtained this method are highly limited knowledge instantaneous between disc rotor pads. Since μ nonlinear function many variables (namely temperature, angular velocity disc), an analytical model appears impractical establish. This work aims implement innovative algorithm, based machine learning, for determining from signals regularly available in enable accurate breaking computation. The proposed consists two main tools. An artificial neural network (ANN) developed approximate unknown that relates input μ, while Kalman filter (KF) implemented estimate real temperature distribution surface constitutes one most important ANN inputs. algorithm has been successfully validated data collected extensive tests racetracks, special sensor setup.