A study on a smart dust network and an analog interface

Abstract: Large sensor networks of very small motes, having sensing, computation, communication and power units, are becoming an active research topic. The major problem in implementing such networks is threefold. Firstly, power consumption and area which are limited by the technology (limitation on minimum size and power consumption of transistors). Secondly, locating the area of event which arises due to requirement of motes without identity. Thirdly, cost factor as the number of motes required would be high. This thesis work was done in two parts, the first part comprising of modeling a power and area efficient smart dust network using a novel algorithm to detect the location of event without giving identity to the motes and also developing an interface for monitoring patient's health in hospitals through such a network. The second part consists of designing an analog front end to generate event in case of abnormalities in signal from human body. The designed front end can also be used for intra body communication systems (body area networks) with operating frequencies of order 10-20 MHz. Body area networks (BAN) is a type of personal area network (PAN), introduced by Zimmerman, using human body as a communication channel for communication between body based devices through capacitive coupling. The major advantage of such communication lies in reducing the burden on RF spectrum. If only one of the body based devices can communicate with the outside world using RF spectrum and all body based devices can communicate with each other forming a BAN, then indirectly all devices can communicate with the outside world with only one of them using the RF spectrum. Power and security are also the inherent advantages in using body as a channel.