Abstract
In recent years, it has been widely believed in the community that the link reliability is strongly related to received signal strength indicator (RSSI) [or signal-to-interference-plus-noise ratio (SINR)] and external interference makes it unpredictable, which is different from the previous understanding that there is no tight relationship between the link reliability and RSSI (or SINR), but multipath fading causes the unpredictability. However, both cannot fully explain why the unpredictability appears in the link state. In this paper, we unravel the following questions: 1) What causes frame losses that are directly related to intermediate link states? 2) Is RSSI or SINR a right criterion to represent the link reliability? 3) Is there a better measure to assess the link reliability? We first configured a testbed for performing a real measurement study to identify the causes of frame losses, and observed that link reliability depends on an intraframe SINR distribution, not a single value of RSSI (or SINR). We also learned that an RSSI value is not always a good indicator to estimate the link state. We then conducted a further investigation on the intraframe SINR distribution and the relationship between the SINR and link reliability with the ns-2 simulator. Based on these results, we finally propose an interference modeling framework for estimating link states in the presence of wireless interferences. We envision that the framework can be used for developing link-aware protocols to achieve their optimal performance in a hostile wireless environment.
Original language | English |
---|---|
Article number | 6200006 |
Pages (from-to) | 272-285 |
Number of pages | 14 |
Journal | IEEE/ACM Transactions on Networking |
Volume | 21 |
Issue number | 1 |
DOIs | |
Publication status | Published - 2013 |
Keywords
- Estimation
- SINR
- frame loss
- intraframe signal-to- interference-plus-noise ratio (SINR) distribution
- link state
- received signal strength indicator (RSSI)
- wireless interference
ASJC Scopus subject areas
- Software
- Computer Science Applications
- Computer Networks and Communications
- Electrical and Electronic Engineering