Solar-powered sensor nets may be the future of border security
High-tech solar-powered sensor networks being developed by researchers at New Mexico State University may someday be part of the U.S. Department of Homeland Security (DHS) defense arsenal, helping to monitor remote regions of the more than 7,000 miles borders shared between the United States, Mexico and Canada.
Hong Huang and Eric Johnson, from the Klipsch School of Electrical and Computer Engineering, are involved in a six-year study funded by DHS that may lead to increased longevity of sensor networks and minimize the environmental impact they pose. At the same time, they may significantly reduce the cost of the devices.
Wireless sensor networks are spatially distributed autonomous devices equipped with sensors that monitor various physical conditions, such as sound, vibration, motion, temperature, etc. They incorporate wireless communications to relay data to a base station and are typically battery powered.
Wireless sensor networks have many uses. Civilian applications include environmental and habitat monitoring, patient-care, manufacturing plant monitoring, home automation and traffic control. Military applications include battlefield surveillance and situation awareness.
DHS uses them to provide unattended, continuous monitoring of remote, inaccessible areas along our borders that are difficult for humans to patrol. By monitoring vibration, for instance, they can detect and differentiate movement that might come from an animal, a human, or a vehicle. They can be dropped by planes and left to do their work where people can’t.
“There are some serious limitations, however, to wireless sensor networks deployed in remote area” said Huang. “Typically, the batteries last only several months up to a year, and the dead batteries pose an environmental hazard along with the hardware that is left in the environment.”
Huang and Johnson are investigating the use of small, palm-sized solar panels that are commercially available to power the sensors and the wireless communications between devices.
“Our calculations show that these small solar panels can provide one-half-watt of power for seven hours under direct sunshine—enough to power several sensors and extend their lifetime for tens of years,” said Huang. “It’s a perfect match, especially here in our border region where we have sunshine 360 days a year.
“The challenge is that a randomly deployed device may not land in direct sunshine and it is not feasible to equip them with auto tracking devices to follow the sun according to time of day and year. There are many dynamics that have to be overcome to effectively and efficiently utilize solar power.”
Additionally, there is a trade-off between the application performance of the sensors and the life-time of the sensor network.
Huang and Johnson are developing solar-aware adaptive algorithms for data collection in sensor networks. These algorithms build on directed-diffusion, a standard data collection algorithm developed at UCLA that does not incorporate solar energy.
Huang and Johnson are designing a smart, adaptive sensor network that will dynamically adjust to the solar power received at various points. They plan to enable the sensor nodes to behave differently, based upon the amount of solar power they have. The solar-rich nodes can take over the burden of relaying data to the base station. Preliminary simulation results show the new method results in energy savings, better performance, and prolonged lifetime of the network.
They are also planning to incorporate into sensor networks passive radio frequency identification (RFID), a technology used by retailers, shippers and the military to track inventory and assets. Passive RFID does not require power to transmit data. A network could be composed of many RFID sensors distributed over a wide area that reflect data to a few solar-powered RFID readers/base stations.
“Because the RFID sensors need no power source, they are very inexpensive and this enables them to be used in widespread applications,” said Huang.
Precision agriculture would be one beneficial use of this technology. Sensors could monitor moisture and soil conditions enabling more efficient irrigation of fields and water savings.
But today’s sensors are costly, said Huang. “If you need 100 sensors per acre, at $100 per sensor, that’s $10,000 per acre, too expensive to be widely used. Incorporating RFID technology can bring the cost down from several cents to several dollars per sensor.”
Huang and Johnson have been investigating sensor net technologies since 2004 for Sandia National Laboratories, Los Alamos National Laboratory and the U.S. Army. They are using network simulation software, radio frequency instrumentation and a sensor network test bed consisting of 100 nodes in their current research. Two doctoral and eight master’s students are assisting with the research. Their current research is being conducted in conjunction with the National Center for Border Security and Immigration, Research, formed last year at the University of Arizona, Tucson.
Linda Fresques
May 5, 2009