A Neuro Fuzzy Approach for Anti-Jerk Control

Comfort is of increasing importance for the decision to buy a car. Modern Diesel engines generate a high engine torque already at low engine speed, which may cause oscillations in the driveline of the vehicle. These oscillations are experienced as unpleasent jerking by the driver. The objective of this article is to present a control concept that prevents the driveline from oscillating so called Anti-Jerk Control.

The development of a controller premises the existence of a profound mathematical model of the process. In this article the identification of the driveline is based on Neuro-Fuzzy models. The structure of Neuro-Fuzzy models allows the consideration of non linear effects, e.g. torsion of the crank shaft, while sustaining the advantages of linear models.

As a measure for the oscillations in the driveline the difference between engine speed and wheel speed is used. Based on a Neuro-Fuzzy model two local controllers are developed. For controller design the root locus technique is used. The controller concept is evaluated on a test car with a diesel engine. The test results will be presented in the paper.