Rate Adaptation for Highly Dynamic Body Area Networks

Body Area Networks (BANs) is a type of wireless sensor networks that enables wearable/implanted sensors to exchange information or collect monitored data. Unlike traditional wireless networks, such as WiFi or bluetooth, where equipments are usually quite stable, sensors are equipped on human body and hence are moved with gesture, posture or human mobility. Such dynamics also make coordination between sensors and the hub more difficult, because the channel quality between nodes could fluctuate significantly within a short period of time. Bit-rate adaptation (i.e., Selection of the modulation and coding scheme) hence becomes one of the most challenging issues needed to be addressed in such a highly-dynamic network. This paper proposes a bit-rate adaptation scheme upon a polling-based MAC protocol. Our design enables real-time channel quality estimation before every data transmission. Body sensors can thus adapt their bit-rates to dynamic channel conditions. Our simulation results show that our rate adaptation scheme maintains the delivery success ratio near 100%. It hence also spends less energy for transmitting data, which further prolongs the lifetime of body sensors.