Vehicle Velocity Estimation by Using Numerical Models

This paper reports the backward Euler, trapezoidal rule, and Taylor expansion models which all estimate vehicle velocity. First-order nonlinear vehicle model of motion is underlying of these three numerical models. They leave out chassis velocity signal from estimations whereas the dedicated state-observer wants such a velocity signal for single state-variable observation. Hence, the observer acquires the force input as well as the velocity output to evaluate merely velocity while the three models obtain only the force input to estimate the same velocity. Thus, the observer proves superfluous and inoperative when the velocity signal turns up missing. Some numerical methods that find solutions of ordinary differential equations (ODEs) are adopted to build the three models. Their fine performance of velocity estimations comparable to the velocity response arising from cruise control system is confirmed by simulation results. Equivalence between speedometer and the three models is feasible and the trapezoidal rule model tends to be more precise.