Declarative querying of sensor networks through automatic service planning

As sensor networks become more and more ubiquitous, engineers face the challenge of making the sensor data accessible to the common person. For example, a building manager might want to be alerted to excess building activity over the weekends, or a safety engineer might want a histogram of vehicle speeds in the parking garage. However, most users don’t want to translate raw sensor data into building activity or vehicle speeds, nor do they want to learn how to write efficient distributed algorithms for wireless, embedded devices.

This talk presents a framework that allows users to program sensor networks with declarative statements such as, “I want a histogram of vehicles speeds in a certain area of the parking garage.” The engine that processes the query will search through existing applications to find useful bits and pieces, and then reassemble them into an application that satisfies the query. If the query cannot be satisfied, it can tell the user which pieces are missing. The query language should also allow the user to specify constraints on the application, such as confidence levels or latency constraints.

This work builds on partial-order planning (POP), first-order logic (FOL), and constraint logic programming (CLP). It adapts previous work on auto-composition of web services such as WSDL and OWL to the event-driven, asynchronous nature of sensor networks.

Speaker Details

Kamin Whitehouse is a PhD student in the TinyOS group at the University of California, Berkeley, where his research focus has been in programming abstractions, sensor field localization, calibration, routing, and collision detection and recovery in wireless networks. He received a M.S. in Computer Science from UC Berkeley in 2002, a B.S. in Electrical Engineering and a BA in both Cognitive Science and Philosophy from Rutgers University in 1999. He has previously done research with AT&T Bell Labs, Intel Research, Xerox PARC, Bellcore and the US Army CECOM.

Date:
Speakers:
Kamin Whitehouse
Affiliation:
MSR Intern/University of California
    • Portrait of Jeff Running

      Jeff Running