DEPEND: Density adaptive power efficient neighbor discovery for wearable body sensors.

Li Bin,Feng Wei,Zhang Lin,Spanos Costas
DOI: https://doi.org/10.1109/CoASE.2013.6653917
2013-01-01
CASE
Abstract:Wearable sensors are designed to monitor the individual user's activities or the social behaviors within a group. Due to the nature of their applications, the wireless transmission of the sensory data are subject to extremely stringent energy constraints. Also, due to user mobility, the network topology is frequently changing. Therefore, the design of efficient communication protocols for wearable sensor networks is technically challenging. As a fundamental building block of the communication protocol, the power-efficient neighbor discovery mechanism is important for the network to function well. We propose an asynchronous neighbor discovery algorithm, namely, DEPEND (Density-based Power Efficient Neighbor Discovery), which requires neither time synchronization nor carrier sensing, and thus is easy to implement on low cost hardware. DEPEND exploits traditional Quorum-based Power-Saving (QPS) scheme, and incorporates an adaptive mechanism to respond to dynamic node density so as to improve and balance the performance of power saving and neighbor discovery. Each DEPEND-enabled network node calculates the local node density and forecasts the probability that new neighbors enter its communication-range in the near future, and adjusts itself to the most suitable QPS mode to discover new neighbors. Analytical and experimental results show that DEPEND is able to improve the neighbor-discovery performance without consuming much more power, thus the discovery performance and the power efficiency are jointly enhanced. © 2013 IEEE.
What problem does this paper attempt to address?