Improving Audio Streaming over Multi-hop Zigbee Networks

Audio streaming over LR-WPAN is different from data communication because such wireless channels are limited in bandwidth and prone to errors. Nevertheless there are several scenarios such as distributed surveillance and emergency response, where lightweight multimedia applications are compelling solutions for pervasive and ubiquitous computing. This paper investigates the feasibility of transmitting voice over ZigBee compliant platforms using commercial off-the-shelf (COTS) hardware. We present an implementation of a ZigBee Push-To-Talk application which allows to investigate the capability of ZigBee protocol for low-rate voice streaming and to analyze important streaming metrics such as packet loss over a multi-hop sensor network, evaluating Unicast and Broadcast addressing. Finally, we analyze the trade-off between audio quality and power consumption comparing different compression algorithms and we draw hard conclusions about specific improvements which would be needed to support higher data rates.