Efficient Conditional-Probability Link Modeling Capturing Temporal Variations in Body Area Networks

Select |


Smith, David; Boulis, Thanassis; Tselishchev, Yuri


Conference Material

The 15th ACM/IEEE International Conference on Modelling, Analysis and Simulation of Wireless and Mobile Systems (MSWIM 2012)

Paphos, Cyprus


This paper presents conditional probabilistic modeling suitable to characterize the temporal variation of links in wireless body area networks (BAN); according to short, medium and long term fading characteristics. The approach captures first and second order statistics appropriately; and using conditional probabilities, predicts what signal power levels can be expected from 10 ms ahead to many seconds into the future in the context of typical autocorrelation and channel coherence times. Hundreds of hours of link measurements, in ``Everyday" mixed activity BAN, are used to generate these conditional models. We show that such modeling has an important effect to higher level simulation results (e.g., packets received at the application layer) for different simulation scenarios. Moreover we show that the model is computationally efficient as its introduction adds only 8% of simulation time on average. It is also shown that short and medium-term conditional modeling can vary considerably from long-term modeling, particularly given lower instantaneous path loss.

Body area networks, channel modeling, conditional probability, time-varying channels



Smith, David; Boulis, Thanassis; Tselishchev, Yuri. Efficient Conditional-Probability Link Modeling Capturing Temporal Variations in Body Area Networks.[Conference Material]. 2012-10-22. <a href="http://hdl.handle.net/102.100.100/99086?index=1" target="_blank">http://hdl.handle.net/102.100.100/99086?index=1</a>

Loading citation data...

Citation counts
(Requires subscription to view)