Driving load estimation with the use of an estimated turbine torque

Abstract

In this paper, a vehicle driving load estimation scheme in the form of a linear state observer is presented. The signals used in the observer are the transmission output speed and driven wheel speed, which are readily available in any vehicle equipped with an automatic transmission. Because the observer requires the turbine torque as input, the turbine torque itself has been estimated using a neural network. The proposed observer has been evaluated using a vehicle simulator in various driving situations considering transmission oil temperature variations, engine power losses, and variation of load conditions. A nonlinear vehicle powertrain model has been used in the development of the vehicle simulator. The effectiveness of the proposed scheme has been tested through experiments.