Modeling and Control with Local Linearizing Nadaraya Watson Regression

TL;DR

This paper introduces a stochastic framework for building black box models of systems using local linearizing Nadaraya Watson regression, demonstrating its application in modeling and controlling a motorcar powertrain.

Contribution

It presents a novel, automated method for constructing black box models with local linearization, applicable in control engineering tasks.

Findings

Framework enables easy, automated black box modeling from observations

Application to motorcar powertrain shows practical effectiveness

Method provides a new approach for system control using local linear models

Abstract

Black box models of technical systems are purely descriptive. They do not explain why a system works the way it does. Thus, black box models are insufficient for some problems. But there are numerous applications, for example, in control engineering, for which a black box model is absolutely sufficient. In this article, we describe a general stochastic framework with which such models can be built easily and fully automated by observation. Furthermore, we give a practical example and show how this framework can be used to model and control a motorcar powertrain.