Underwater Sensor System Could Protect Reservoirs, Drinking Water

"The importance of developing a network sensor technology for operation in liquid environments has recently been highlighted in reports detailing the chemical slurry of antibiotics, estrogen-type hormones, insecticides, nicotine and other chemicals in the rivers of industrialized countries," says Dr. Craig A. Grimes, associate professor of electrical engineering and materials science and engineering. "However, analysis is still done by physically collecting samples and analyzing them back in the laboratory."

Monitoring of rivers downstream from sewage treatment plants, large city water supplies, or the composition of a local pond must all be done by hand. This expensive, time-consuming and sometimes dangerous practice is always time delayed and may miss short duration episodes of pollution or contaminants. Continuous, in-place monitoring would be the easiest, most timely and least expensive way to track changes in bodies of water.

However, underwater monitoring is hampered because water interferes with the radio transfer of information, the most common method used to transfer information in the air. The researchers, who include Grimes; Xiping Yang, William R. Dreschel, Kefeng Zeng and Casey S. Mungle, graduate students, electrical engineering, Penn State; and Keat G. Ong at SenTech Corporation, State College, Pa., looked at a hierarchical, acoustic method to transfer the information from the sensors to the person monitoring the water.

The researchers are looking at systems that can monitor temperature, salinity, acidity and specific chemicals. Some of the same researchers, in collaboration with Dr. Michael Pishko, associate professor chemical engineering and material science at Penn State, are working on an inexpensive, disposable sensor for ricin, the highly poisonous protein found in castor beans and thought to be a potential terrorism agent. Sensors also exist for other harmful chemicals.

In the aqueous sensor network system an uplink node floats on the water's surface and transfers the aqueous network data from the water to the air, where it is received by the command computer.

Beneath the surface, layers of nodes/sensors monitor the water and pass the information along to the uplink. Sending a message from the farthest node direct to the uplink underwater is not possible because of the way water decreases the strength of the acoustic signal, so the researchers use a node-to-node multi-hop information transfer system.

"Node-to-node communication enables wide-area coverage using modest node power levels making practical long-term monitoring," Grimes reported in a paper in the journal Sensors.

After the network of nodes is deployed, floating anchored in place in the water, the system must set up an identification tree. The uplink node broadcasts a signal containing its identity. Every node that receives that broadcast marks the uplink node as its parent node. These nodes then broadcast a signal. Every node that receives that signal, and has not yet identified a parent node, will record the signaling node as its parent and then broadcast to even more distant nodes. A cascade of parent nodes eventually covers the entire system.

Periodically, the network sends data through the system. Each node sends its sensor data to its parent node. That node sends the received data and its own data to its parent node until all the data are received by the uplink node, which converts the signal from acoustic to radio frequency and sends the information through the air to the command, or central, computer for display and evaluation.

The host node stores the sensor data from all the nodes in its memory preserving the identity of the node that produced the data so that water-monitoring personnel can track unusual readings or contaminants to their source location.

The researchers designed the nodes so that the chemical sensors are immersed in water separate from the communication electronics, making it easy to change the sensors on the nodes without having to alter the signaling network.

The National Science Foundation supported this work.