Wireless sensor networks represent a disruptive technology likely to have a great impact on both society and research throughout this decade. Programming sensor networks is a difficult and error prone task as they often exhibit continuous reconfiguration caused by environmental factors and energy depletion, the introduction of new sensors, and the presence of various forms of mobility internal and external to the network. This project is developing a flexible software infrastructure that allows the programmer to treat the sensor network as a shared, evolving, geographically distributed, general-purpose computing milieu. This infrastructure achieves an unprecedented level of flexibility, fluidity, and adaptability by structuring applications in terms of a community of coordinating mobile agents. Agents are able to enter and exit the sensor network and to construct and evolve complex applications by exploiting new forms of code and state mobility and dynamic composition. They are able to maintain the integrity of the application and its relation to the physical environment being monitored by exploiting new coordination techniques that include a spatiotemporal dimension. A distinguishing feature of the proposed software infrastructure is its unified treatment of all forms of mobility encountered in sensor network applications: movement of entities in the physical environment, movement of the user, movement of code within the network and across network boundaries, and even the movement of sensors making up the network. This project is creating the very first agent-based, software infrastructure for the development of flexible sensor network applications. It allows programmers to design applications that can be deployed when needed and as needed, can share network resources, can limit their network usage only to regions of interest, can adapt to network and environmental conditions, and can reduce energy consumption.

This research was made possible by NSF Grant CNS-0520220. It was conducted within the Department of Computer Science and Engineering at Washington University in St. Louis.


The following research projects were made possible by this grant.


The following are the researchers supported by this grant.