No abstract

In this research work, three types of deployment models, namely regular grid, uniform random, and actual town models are compared regarding communication performance prediction of multi-hop wireless smart meter networks. Multi-hop data transmission simulation and analysis are carried out to find the most suitable type of generic as well as realistic meter deployment model that would facilitate planning on access point location for each town area. We introduce a spatial analysis index that quantifies the randomness and uniformity of spatial deployment. In addition, our analytical investigation using gamma statistical distribution shows that such a random model with a larger gamma shape parameter is the most suitable generic and realistic meter deployment model for multi-hop data transmission performance evaluation.

Performance of the ‘transmission reserve table’ (TRT) buffer, which the authors recently proposed, is theoretically investigated. The buffer's triage ranking queuing was designed for resource‐limited control networks that are possibly exposed to a rush of control commands. Our previous study analyzed the performance of the TRT buffer by means of computer simulation. In this paper, we derive a stochastic analytical model for the TRT buffer and verify the model in comparison with computer simulations. Theoretical calculation demonstrates that the triage ranking of the TRT plays a crucial role in keeping the real‐timeness of the control network even in a rush of commands by our triage ranking mechanism. © 2015 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.

In this research work, three types of deployment models of multi-hop wireless smart meter networks have been compared. Three types of models, namely, regular grid, uniform random, and actual town models were constructed. Multi-hop data transmission simulations were carried out to obtain packet delivery rates and data transmission times from each smart meter to an access point. The variation ranges of packet delivery rates for the random and actual models were approximately 3 times larger than that for the regular models. The data transmission times for the actual and random models were relatively worse than those for the regular models. We introduced two spatial analysis indexes and found that these indexes of the actual models are similar to those of the random models rather than the regular models. These indexes are possibly used to estimate the communication performance of the smart meter network in an actual deployment.

This paper studies an architecture of multi-server Web industrial facility monitoring systems using only commercial-off-the-shelf (COTS) Web browser's existing functionality. First we introduce our original method of background change-of-state (CoS) communication while the user interface is focusing to a currently-connected foreground Web server. The method makes it possible for currently-notconnected background Web servers to push a CoS message to the COTS browser. Next, we show our automatic recovery method from COTS browser's communication errors and our experiment results of COTS browsers' long-term memory leaking characteristics.