Sensor Network Interoperability and Reconfiguration through Mobile Agents
Open Access
- Author:
- Jean, Evens
- Graduate Program:
- Computer Science and Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- February 25, 2011
- Committee Members:
- Dr Ali Hurson, Dissertation Advisor/Co-Advisor
Dr Ali Hurson, Committee Chair/Co-Chair
Dr Robert T Collins, Committee Chair/Co-Chair
Vijaykrishnan Narayanan, Committee Member
Ling Rothrock, Committee Member
Dr Yu (Cathy) Jiao, Committee Member - Keywords:
- mobile agents
sensor network
FPGA - Abstract:
- Sensor networks generally consist of small devices deployed in an area to perform a task through coordination and communication. The current paradigm in sensor network promotes isolated networks with statically tasked nodes; this hinders the potential for applications to harness the power of fused data from multiple networks. Furthermore, while individual networks subsist in a dynamic environment, the tasks of the nodes are generally static and cannot adapt to changes in application requirements. To address these issues, this dissertation aims at rendering sensor networks interoperable and reconfigurable through reliance on mobile agent technology. The proposed approach intends to leverage data collected from individual networks to increase the set of knowledge available to applications. Our research has led to the introduction of a Reconfigurable and Interoperable Sensor Network (RISN) architecture that relies on mobile agents to achieve the stated goal. RISN provides network interoperability, and supports dynamic tasking of nodes through agent migration. The system further relies on Field Programmable Gate Arrays (FPGA) to handle potential issues associated with the execution time and latency of results by allowing applications to take advantage of hardware accelerators. The use of agent technology introduces security risks to the system; as such, we also investigated the security issues that plague agent-based systems. As the use of agents in practical applications is primarily limited by the security concerns of hosts, our investigation resulted in the introduction of a Distributive and Adaptive Security-Monitoring through Agent Collaboration (DASAC) to address such concerns. DASAC classifies agents based on their execution patterns on current and prior hosts in order to thwart attacks thereby abating the security concerns of hosts. Lastly, as a proof- of-concept, we discuss the implementation of a distributed target-tracking application that relies on the interoperability and reconfiguration of networks that RISN affords to applications, in order to continuously maintain the location of object of interest despite the heterogeneity of the network, and limited views of any one sensor.