Drs. Rodger Ziemer and Mark Wickert, ECE
Wireless sensor networks will have a significant impact on our lives in the 21st century1. Networks of miniaturized inexpensive devices will have the capability of measuring a myriad of quantities, including light, vibrations, humidity, heat, acoustics, etc., then performing initial processing on these measured quantities, forwarding them via electromagnetic communication links to central decision-making nodes, and finally implementing a desired course of action based on the sensed information. It is estimated that the transfer of one bit of information costs roughly 10 to 100 times more in terms of energy consumption than to store and process it, making communications a big challenge in sensor networks.
The ideal sensor network possesses several attributes including ease of setup, low energy consumption, low probability of intercept, resistance to jamming, the ability to reconfigure if one or more nodes become inoperable, and some measure of security. These attributes point to wireless self-con-figuring networks with the ability to remain undetected and, once detected, to resist jamming by a hostile party. Furthermore, the nodes, or communicating entities, of these networks should be small, lightweight, and give long service with limited energy sources.
The focus of this research is to evaluate the suitability of two competing modulation techniques for application to sensor networks - direct-sequence spread-spectrum and ultra-wideband pulse position modulation. Ultra-wideband is motivated by recent spectral usage changes made by the Federal Communications Commission. A general block diagram of the system being studied is given in the first Figure below. The physical layer modeling of interest is shown inside the dash-dot line.
Packet error rate results were obtained for a path through the geometrical configuration of sensors shown in the second Figure below. The results characterize direct-sequence spread-spectrum and ultra-wideband modulation schemes used to route sensor data from a remote sensor to a collection point via repeater nodes. The number of repeater nodes was varied along a path through a configuration of other emitting sensors. The two wideband communications schemes considered on the basis of packet error rate, direct-sequence spread-spectrum and ultra-wideband, were found to per-form comparably for the parameter values chosen. Thus, initial results indicate that the choice of modulation scheme is basically neutral. Currently being considered are other sensor configurations, modem parameter choices, and the use of coding to improve link performance.
The recent work highlighted above is sponsored by NISSC, and is closely related to earlier work done by the investigators on satellite and wireless communications networks. The earlier work was sponsored by the Office of Naval Research from 1986 to 1998, which resulted in over $1.3M of research funding and involved two other investigators in Electrical and Computer Engineering and Computer Science, Drs. Charlie Wang and Edward Chow.
Dr. Ziemer joined the University of Colorado at Colorado Springs (UCCS) in January 1984 as Professor and Chairman of the Electrical and Computer Engineering Department, and is a prolific author, including several books listed below.