Measurements and characterization of link quality metrics in energy constrained IEEE 802.15.4

[Summary form only given]. As more and more low-cost high-quality devices appear on the market and new applications emerge every day, short-range wireless personal area networks (WPANs), both low and high-data-rate, are on the horizon. Two major efforts of IEEE are underway to enhance the development of WPANs. One is the specification of IEEE 802.15.3a, also known as ultra wideband (UWB), used in high-rate WPANs. The other is the specification of IEEE 802.15.4 for low-rate WPANs (LR-WPANs). When deploying LR-WPAN devices into a target field, information about how sensors perform at various distances and orientations relative to each other will make optimal placement easier. If each node is placed to maximize its performance in the network, the lifetime and quality of service for the entire network will increase, therefore an understanding of the wireless channel-operating environment is necessary. In this work, we evaluate through measurements, the basic characteristics of the communication channel using IEEE 802.15.4 radio devices. The experiments were conducted on the esplanade at the Engineering building. It consisted in capture the spatial radio channel properties due to stationary structures and obstacles, and the resulting findings will serve as the basis of an eventual methodology for predicting radio channel quality given an arbitrary layout. A simple transmission scheme was designed, where device-1 (receiver) acted as an access point (AP) and device-2 (transmitter) acted as a sensor placed at a location of interest. The esplanade was divided into 26×32 1m-by-1m grids, and device-2 was sequentially placed at each grid point (x,y), for one measurement. The RSSI and PER of each received packet were saved. Average RSSI and PER metrics presented for each link measurement were calculated over 3000 transmitted packets. The results clearly show the anisotropic nature of all metrics. Specifically, radio links with the same line-of-sight distance did not necessarily have similar average RSSI, or PER. As the superposition of multipath components arriving from different paths can often result in higher energy detected by the receiver (RSSI), but the resulting signal can easily be distorted from the original waveform given the phase differences of those components. About 90% of the links had a PER less than 0.005, and the worst link observed got a PER of 0.054. Links with 0.85 or higher packet reception ratios are considered good links. Links with 0.15 or less packet reception ratios are considered bad links. For one link, however, no obvious blockings but only few clutters were observed; the lower link qualities are thus attributed to multipath effects due to surrounding reflectors. Radio links between the sensors radios exhibited different path losses at different locations. The path loss was not only a function of path distance alone but also showed complex dependency on blocking and multipath effects caused by the surrounding obstacles. In general, good quality was observed for the links; most PERs were less than 0.01, which is consistent with previous works. The experiments in this paper are preliminary studies toward exploring a methodology to predict radio performance at any location.