Researchers to Develop Sensor Networks that 'Talk' and 'Think'

AMHERST, Mass., Oct. 23 -- With virtually every segment of modern life -- from healthcare and manufacturing to homeland security -- relying increasingly on remote sensing technology, a group of researchers is working to create a new generation of intelligent and networked sensors that can communicate with each other and use the shared data to make decisions.

An interdisciplinary team from the University of Massachusetts (UMass) Amherst and Boston University have been awarded a $2.47-million, five-year grant from the National Science Foundation to develop a self-contained sensor network that can adapt to its environment based on the data it gathers.

Mechanical and industrial engineering professors Robert Gao and Abhi Deshmukh, and electrical and computer engineering professor Weibo Gong make up the UMass team; the BU team is led by former UMass Amherst faculty member Christos G. Cassandras and includes David A. Castanon, Ionannis Paschalidis and John Baillieul.

According to Deshmukh, the interdisciplinary team is taking a bold course by simultaneously developing the hardware for the sensors and creating the network for the devices to communicate and reach decisions about the health of the system.

Their initial efforts will focus on developing a manufacturing model to demonstrate the sensor technology. Gao says a low-power sensor network will monitor the condition of the machinery. But the potential uses for "smart" sensor networks go well beyond the factory floor, according to the researchers.

"There are applications for this technology in almost every area of society, such as manufacturing, intelligent transportation systems, weather and radar surveillance and detecting biochemical agents," says Deshmukh. "Almost any future system will need an integrated sensor system."

Gong said, "Wireless sensor networks need new algorithms and analysis methods for large, complex engineering systems and pose challenges to fundamental research. We are very excited to work on these issues."

By linking the various sensors into a communication grid, says Gao, the network will be able to compensate for faulty readings or poor quality information coming from individual devices.

Gaowill also be concentrating on building an energy-efficient sensor that can extend its battery life by self-regulating its operational and communication activities. "People want to put sensors everywhere, but they’re not much good if you have to change the batteries every week," he says.

The BU team will work on various network control issues and experimentally link several laptop computers with small robots outfitted with sensors to create a localized sensor network.

For more information, visit: www.ecs.umass.edu