Discharge Prediction of Rechargeable Batteries with Neural Networks

Authors:
Olivier Gérard;JEan-Noël Patillon;Florence D'Alché-Buc
Affiliations:
Laboratoires d'Electronique Philips S.A.S. (LEP), 22 Avenue Descartes, B.P. 15, 94453 Limeil-Brévannes, France (Correspd. gerard@lep.research.philips.com);Laboratoires d'Electronique Philips S.A.S. (LEP), 22 Avenue Descartes, B.P. 15, 94453 Limeil-Brévannes, France;Laboratoires d'Electronique Philips S.A.S. (LEP), 22 Avenue Descartes, B.P. 15, 94453 Limeil-Brévannes, France
Venue:
Integrated Computer-Aided Engineering
Year:
1999

Citing 7
Cited 1

Programming a massively parallel, computation universal system: Static behavior

AIP Conference Proceedings 151 on Neural Networks for Computing
A self-optimizing, nonsymmetrical neural net for content addressable memory and pattern recognition

Physica D
Multilayer feedforward networks are universal approximators

Neural Networks
Introduction to the theory of neural computation

Introduction to the theory of neural computation
Neural networks for control

Neural networks for control
Intrinsic and Parallel Performances of the OWE Neural Network Architecture

ICANN 96 Proceedings of the 1996 International Conference on Artificial Neural Networks
Control of nonlinear dynamical systems using neural networks. II. Observability, identification, and control

IEEE Transactions on Neural Networks

Improved spiking neural networks for EEG classification and epilepsy and seizure detection

Integrated Computer-Aided Engineering

Quantified Score

Hi-index	0.00

Visualization

Abstract

This article presents an original method to accurately predict the end of discharge of rechargeable batteries inserted in portable electronic equipments. The proposed method is based on two neural networks organized in a master-slave relation. A prediction accuracy of 3% (18 minutes) is reached. A further improvement of the system is introduced by adapting on-line another neural network to the actual battery currently in use. This adaptive method reduces the average error to 10 minutes. Results are promising and implementation, carried out in a portable multimeter prototype, only requires a small amount of the computing power already available inside most portable equipments.