Extended Markov Tracking with Ensemble Actions


Rabinovich, Zinovi and Jennings, Nick (2009) Extended Markov Tracking with Ensemble Actions. In, International Workshop on Hybrid Control of Autonomous Systems

Download

[img] PDF - Submitted Version
Download (139Kb)

Description/Abstract

In this paper we extend the control methodology based on Extended Markov Tracking (EMT) by providing the control algorithm with capabilities to calibrate and even partially reconstruct the environment's model. This enables us to resolve the problem of performance deterioration due to model incoherence, a negative problem in all model based control methods. The new algorithm, Ensemble Actions EMT (EA-EMT), utilises the initial environment model as a library of state transition functions and applies a variation of prediction with experts to assemble and calibrate a revised model. By so doing, this is the first control algorithm that enables on-line adaptation within the Dynamics Based Control (DBC) framework. In our experiments, we performed a range of tests with increasing model incoherence induced by three types of exogenous environment perturbations: catastrophic, periodic and deviating. The results show that EA-EMT resolved model incoherence and significantly outperformed the best currently available DBC solution by up to $95\%$.

Item Type: Conference or Workshop Item (Paper)
Divisions: Faculty of Physical Sciences and Engineering > Electronics and Computer Science > Agents, Interactions & Complexity
ePrint ID: 267337
Date Deposited: 07 May 2009 11:55
Last Modified: 27 Mar 2014 20:13
Further Information:Google Scholar
URI: http://eprints.soton.ac.uk/id/eprint/267337

Actions (login required)

View Item View Item