Differentiating emergency voice traffic in indoor Wireless Activity Monitoring Network

We investigate QoS provisioning in a Wireless Activity Monitoring Network(WAMN) when data traffic is a mixture of emergency short voice commands and 3D accelerometer signals for remote monitoring of a group of elderly people living in a health care building. Nodes equipped with acoustic and accelerometer sensors form a multi hop wireless sensor network. The network carries multi class traffic to the sink over a service differentiation protocol for processing. Utilization of limited resources is critical for feasible quality of service. We analyze network performance trade-offs and elaborate on optimal buffer length, MAC retransmission count and transmit power for best achievable voice quality at the sink. Based on the results of extensive simulations, we conclude with some remarks on the tuning off network resources for data segment reliability and voice quality in terms of a novel R-factor metric for the given scenario.