A generic rules-based inference engine to track mobiles with secure localization techniques

Project: PHD

Description

Motivations

Surveillance of some categories of individuals with electronic means could improve public security. Some because they represent a potential threat for the society (recidivist sexual offenders, terrorists...), some because they are vulnerable (elder people, patients...) and eventually some because of the risks linked to their duty in a public organisation (soldiers, firemen, policemen...).

Some objects (weapons, precious material, dangerous substances...) should be tracked as their loss could threaten public security or lead to heavy financial prejudices for the federal state or the regions.

Description

The present thesis covers the analysis, the design and the prototyping of a generic rules-based inference engine that can address the needs of different surveillance applications. Those applications, allowing tracking people or objects, are aimed to maintain or improve public security.

The different 'instances' of this generic software could be used by control centres monitoring prisoners or sex offenders in probation period, by security services command posts, by military headquarters...

The development of geo localisation technologies allows a quasi permanent and ubiquitous surveillance and thus addresses the stringent requirements of some categories of those applications (sex offenders' surveillance).

This generic software will offer interfaces to different geo localisation technologies (location servers...), to equipments bound to individuals and objects to be tracked (electronic bracelets, implants, RFID chips...) et to terminals used by the end-users of those applications: the 'controllers'.

One must address the potential risks that such applications represent for privacy and personal freedom. Therefore, this software will include security measures aimed to prevent, or at least detect, the potential abuses from the 'controllers'.

Methods and tools

The analysis and the design will be based on UML (Unified Modelling Language).

The requirements for the different applications will be collected at the Belgian justice, homeland and defence federal public services.

This generic rules-based inference engine will be developed and deployed on open source platforms (operating system, database management system, integrated development environment...).

It will implement artificial intelligence techniques: A set of adaptive rules that will be easily applied to the different surveillance applications and will allow detecting 'abnormal' situations: A tracked individual or object is not located at the expected location or does no longer emit signals, a 'controller' is performing a suspect query... Moreover, artificial intelligence techniques could also be used to improve the motion model of the mobiles to be tracked.

It will also make use of statistical methods as one has to determine the position of a mobile at a given time from localization data collected from different sources at different times with different levels of accuracy (data sources: GPS, triangulation in a cellular network, Argos...). So, the system will not provide the position of a mobile but rather the likelihood that a mobile is positioned at a certain location with a certain confidence interval.

At last, it will rely upon information security protocols, mitigating the risks of specific attacks threatening geo localization systems ('wormholes'...) and guaranteeing the control of the 'controllers'.
StatusCurtailed
Effective start/end date1/09/0415/09/07

Keywords

  • security
  • Location-based services
  • tracking