Steering and Lateral Motorcycle Dynamics Estimation: Validation of Luenberger LPV Observer Approach

Abstract

In this paper, we proposed a realistic validation of the nonlinear Luenberger observer initially introduced in one of our recent work. The validation is performed by means of a commercial motorcycle simulator. The previous paper has introduced the design of a nonlinear Luenberger observer to simultaneously estimate both the motorcycle lateral dynamics and the rider action. The vehicle lateral motion is described with a two-body linear model which takes into account the time-varying longitudinal velocity. Then, the Takagi–Sugeno approach combined with the Lyapunov theory, linear matrix inequalities tools and $L_2$ -gain property are used to demonstrate the observer's convergence through an input-to-error stability. Finally, the effectiveness and the robustness of the observer are illustrated on three different realistic riding scenarios by means of co-simulation of Matlab/Simulink and the well-known multibody simulator BikeSim. Finally, the estimation performances are compared to previous works.