Security - network security, authentication, access control, trust management
Management - policy based management, self management, autonomic management
Ubiquitous and PervasiveComputing - ad hoc sensor networks, body sensor networks

This project concerns the design of a suitable key management and membership management framework for a Personal Area Network (PAN) involving a small number of sensors worn by patients (also includes implantable sensors). It also involves the design of a suitable approach to securely associate two or more devices. In a medical monitoring system, a patient wears several sensors to monitor his physiological data, the challenge is to securely associate the sensors with a particular patient.  Furthermore, key management is important as it ensures that medical data are transmitted in a secure manner (data confidentiality). The project should conduct a survey of various key management algorithms for wireless sensor networks in the literature. It should either adopt a suitable algorithm for use in a PAN or design a better scheme based on the limitation of existing algorithms. A prototype implementation is necessary using Imperial's BSN motes. In addition, a simulation of the communication overhead and costs incurred by the key management would be needed as part of the evaluation.

Future e-Science and e-Health applications will involve mobile users, possibly with on-body sensors interacting with a ubiquitous computing environment which detects their activity, current context and adapts accordingly. However, the promise of such ubiquitous computing environments will not be realised unless these systems can effectively "disappear"; and for this they need to become autonomous by managing their own evolution and configuration changes without explicit user or administrator action. This project will develop the architecture, tools and techniques which permit these environments to become self-managing. To provide self-management at varying levels (for individual devices, for simple body-area or home-area networks, as well as large-scale network infrastructures) we advocate the concept of a self-managed cell (SMC) as the basic architectural pattern at both local and integrated levels. We will define, prototype and evaluate architectures based on the SMC pattern and their use in e-Health applications. To this end we will: define and implement the core SMC pattern in terms of the monitoring, service-discovery, context and policy-control services required for basic adaptation mechanisms, investigate how SMCs can be dynamically structured into larger structures and specialised SMCs and their interactions for two e-Health application scenarios.