A multi-agent framework for developing complex software
A multi-agent framework for developing complex software
Human Biometrics, with gait as an example, has been around as a science for several years. Applications of the gait biometric can be found in the fields of medical diagnosis, physical therapy and sports. However, it is only in recent times that its use for human iden tification has become practical for real life and real time applications due to the increase in commercially available computing power. The availability of a suitably powerful computing platform is only one aspect of building such system, there still remains the question of how to realise its software architecture. 'With out a viable solution to this problem the application and use of gait identification in a real life scenario will remain in its infancy. Such a system is complex in two respects. One is the complexity in the computations of its individual components and the other is the complexity in building the layout of these components and communications between them. Additionally there is the difficult task of efficiently maintaining such a system. If the frame work that holds the components is not flexible enough, adding new components will be an extremely expensive and difficult task. Therefore, this work focuses upon the creation of a general-purpose, multi-agent framework. As such, the primary goal of this architecture is to reduce system complexity allowing ease of maintenance and expansion, while also allowing distribution of data processing. This framework is entirely general in its application and can be used for holding any soft ware system capable of distributed execution. By using gait identification in this work as an example application, the framework demonstrates its capabilities and thus the advantages of using it.
University of Southampton
Gordon, Layla
6e6541cb-70d5-4285-b1a1-9bdbf9248e9b
2007
Gordon, Layla
6e6541cb-70d5-4285-b1a1-9bdbf9248e9b
Gordon, Layla
(2007)
A multi-agent framework for developing complex software.
University of Southampton, Doctoral Thesis.
Record type:
Thesis
(Doctoral)
Abstract
Human Biometrics, with gait as an example, has been around as a science for several years. Applications of the gait biometric can be found in the fields of medical diagnosis, physical therapy and sports. However, it is only in recent times that its use for human iden tification has become practical for real life and real time applications due to the increase in commercially available computing power. The availability of a suitably powerful computing platform is only one aspect of building such system, there still remains the question of how to realise its software architecture. 'With out a viable solution to this problem the application and use of gait identification in a real life scenario will remain in its infancy. Such a system is complex in two respects. One is the complexity in the computations of its individual components and the other is the complexity in building the layout of these components and communications between them. Additionally there is the difficult task of efficiently maintaining such a system. If the frame work that holds the components is not flexible enough, adding new components will be an extremely expensive and difficult task. Therefore, this work focuses upon the creation of a general-purpose, multi-agent framework. As such, the primary goal of this architecture is to reduce system complexity allowing ease of maintenance and expansion, while also allowing distribution of data processing. This framework is entirely general in its application and can be used for holding any soft ware system capable of distributed execution. By using gait identification in this work as an example application, the framework demonstrates its capabilities and thus the advantages of using it.
Text
1069862.pdf
- Version of Record
More information
Published date: 2007
Identifiers
Local EPrints ID: 466230
URI: http://eprints.soton.ac.uk/id/eprint/466230
PURE UUID: 37fb0448-da96-4c76-aa9b-6501c90bb841
Catalogue record
Date deposited: 05 Jul 2022 04:52
Last modified: 16 Mar 2024 20:34
Export record
Contributors
Author:
Layla Gordon
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics