Vehicle Sideslip Angle Estimation Through Neural Networks: Application to Numerical Data

Aim of this paper is to evaluate the possibility of estimating the vehicle sideslip angle through a non-structured algorithm based on neural networks. Results reported are relevant to a numerical investigation of the network performance which can be regarded as preliminary stage for the application on a real vehicle. A numerical model is used to describe the vehicle dynamics and to generate the inputs for the neural network; with an appropriate set of manoeuvres for network training the non-structured algorithm provides reliable results when applied to a complete series of handling manoeuvres carried out with different tire-road friction coefficients.