AbstractsBiology & Animal Science

Improvement of a three-tier wireless sensor network for environment monitoring

by Xu Wang




Institution: Kansas State University
Department: Department of Biological & Agricultural Engineering
Degree: PhD
Year: 2014
Keywords: Wireless sensor network; Engineering, Agricultural (0539)
Record ID: 2031852
Full text PDF: http://hdl.handle.net/2097/18355


Abstract

A three-tier wireless sensor network (WSN) was developed and deployed to remotely monitor suspended sediment concentration and stream velocity in real-time. Two years of field experiments have demonstrated the achievement of such capabilities. But several weak points emerged and required essential performance improvement and additional research on the radio propagation mechanism within the original three-tier WSN. In the original three-tier WSN, long time delay, potential data loss, and limited network throughput all restricted the network transmission performance. Upon the above issues, the transmission delay was reduced through shortening the raw data storage buffer and the data packet length; the data loss rate was decreased by adopting a mechanism using semaphores and adding feedback after data transmission; the network throughput was enlarged through the event- and time-driven scheduling method. In order to find a long-range wireless transmission method as an alternative to the commercial cellular service used in the original WSN, a central station using meteor burst communication (MBC) technology was developed and deployed. During an 8-month field test, it was capable of performing long distance communication with a low data loss rate and transmission error rate. But due to unstable availability of the meteor trails, the MBC network throughput was constrained. To reduce in-situ maintenance, over-the-air programming was implemented. Thus, programs running in the central station and the gateway station can be updated remotely. To investigate the radio propagation in densely vegetative areas, a 2.4 GHz radio propagation path loss model was derived to predict the short-range path loss from the path loss in the open area and the path loss due to dense vegetation. In addition, field experiments demonstrated that ambient air temperature, relative humidity, and heavy rainfall could also affect wireless signal strength.