fusion年会10.1.1.10.7251.pdfVIP

Multisensor Data Fusion in Distributed Sensor Networks Using Mobile Agents Hairong Qi, Xiaoling Wang S. Sitharama Iyengar Krishnendu Chakrabarty ECE, 319 Ferris Hall CS, 298 Coates Hall ECE, 130 Hudson Hall University of Tennessee Louisiana State University Duke University Knoxville, TN 37996 Baton Rouge, LA 70803 Durham, NC 27708 hqi,xwang1@ iyengar@ krish@ Abstract – We describe the deployment of mobile agent for data fusion tasks as an alternative to the traditional in Distributed Sensor Networks (DSNs) to form an im- client/server model. pr oved infrastructure for multisensor data fusion. Com- The outline of this paper is as follows: Sec. 2 de- par ed with the traditional client/server par adigm, mo- scribes the mobile agent computing model and its ad- bile agent adopts a new computing model: data stay vantages over client/server model. Sec. 3 presents the at the local site, while the execution code is moved to design and implementation of mobile-agent-based DSN the data sites. Mobile-agent-based DSN (MADSN) saves (MADSN). Sec. 4 discusses different fusion algorithms the network bandwidth and pr ovides an effective way to including both centralized and distributed design. Sec. 5 overcome network latency. It is used to integrate pr e- uses target classification as an example to show the effec- pr ocessed data located at local sensor nodes. Progres- tiveness of MADSN in distributed sensor fusion. sive accuracy can be achieve during the agent migration. We take target classification as an example to show h