Modelling human control behaviour with a Markov-chain switched bank of control laws

Abstract

A probabilistic model of human control behaviour is described. It assumes that human behaviour can be represented by switching among a number of relatively simple behaviours. The model structure is closely related to the Hidden Markov Models (HMMs) commonly used for speech recognition. An HMM with context-dependent transition functions switching between linear control laws is identified from experimental data. The applicability of the approach is demonstrated in a pitch control task for a simplified helicopter model.