Print

Capsuto, Benji

Electrical and Computer Engineering

2006

MS

Inconsistencies in wireless channels due to multipath and interference are expected in dynamic environments such as for mobile systems or industrial and office settings. In this work, we show that even static deployments for both indoor and outdoor applications are impacted by variable radio signals. Furthermore we explore the impact of such variations on link quality and localization methods. Two systems were deployed at the University of Vermont; an outdoor system to characterize the effects of weather phenomenon on sensor network communication links, and an indoor system to study frequency-selective multipath fading. Outdoors, three links were monitored, each at different frequencies in the 2.4 GHz ISM band. While it is generally accepted that rain fade does not affect line of. site (LoS) wireless systems below 10 GHz, our results indicate this does not imply that these links were completely impervious to weather related issues. The indoor system was deployed to illustrate the spatial variability in signal strength for a seemingly homogeneous environment. Sensor nodes experience particularly severe adverse multipath fading effects due to their close vicinity to the ground. This environment was characterized using the log-normal shadow model on a channel, location and global scale. The resulting models were used with a multilateration localization method to position the nodes. There are three key results of our work. First, sensor nodes deployed outdoors will experience significant radio signal fades due to precipitation events. These fades will impact the sensor network's ability to maintain a communication link. Second, sensor nodes experience particularly severe adverse multipath fading effects due to their close vicinity to the ground, as illustrated in the indoor deployment. Finally, we've shown that random fluctuations in RSSI wiIl produce wild predictions in node location unless a sufficient number of channels are used.