Model Analysis of Electrically Driven Vehicles by Means of Unknown Input Observers

Abstract

The article describes a method to analyze the powertrain operation of electrically driven vehicles in cases of insufficient information (i.e., unknown control algorithms, no torque measurements during vehicle tests). The method implies mathematical modeling with involvement of so-called unknown input observers. A variant of such an observer is proposed. Using that observer, studies of a hybrid vehicle and a pure electric vehicle are performed. The models with torque observers simulated tests of said vehicles conducted on a chassis dynamometer and on roads. For the hybrid vehicle, operating regimes of main powertrain components were identified. For the electric vehicle, the identification revealed a coordinated operation of regenerative braking and mechanical braking. Adequacy of the modeling, including identification of the unmeasured torques, was verified through a comparison with experimental data.