Chernoff information path loss in wireless relay sensor networks

The author studies wireless relay sensor networks (WRSNs) with randomly deployed sensors for detection of a phenomenon-of-interest (POI). The existing literature in this area has two main drawbacks. Firstly, little has been done to understand whether it is beneficial to relay in WRSNs and what would be the most suitable relaying mechanism such as decode-and-forward (DF) or amplify-and-forward (AF). Secondly, oversimplification in sensing and communication channel models is routinely enforced as a result of the dimensionality and complexity of the `whole-network` analytical setting used in the existing works. To alleviate the aforementioned constraints, the author adopts a `per-path` approach by looking into one typical information pipeline connecting the POI to the destination through either single-hop (SH) or multi-hop (MH). The author investigates the corresponding Chernoff information path loss (CIPL) as the sensor attempts to send 1-bit information about the POI to the remote data sink. The CIPL quantifies the detection reliability against the distance of separation between the original sensor and the data sink, which parallels the classic concept of power path loss (PPL) in wireless propagation channels. The notion of CIPL would provide useful insight on the problems of when and how to relay in WRSNs.