Efficient Data Collection Over Multiple Access Wireless Sensors Network
Alejandro Cohen,Asaf Cohen,Omer Gurewitz
DOI: https://doi.org/10.48550/arXiv.1712.03314
2018-10-16
Abstract:Data collection in Wireless Sensor Networks (WSN) draws significant attention, due to emerging interest in technologies raging from Internet of Things (IoT) networks to simple "Presence" applications, which identify the status of the devices (active or inactive). Numerous Medium Access Control (MAC) protocols for WSN, which can address the challenge of data collection in dense networks, were suggested over the years. Most of these protocols utilize the traditional layering approach, in which the MAC layer is unaware of the encapsulated packet payload, and therefore there is no connection between the data collected, the physical layer and the signaling mechanisms. Nonetheless, in many of the applications that intend to utilize such protocols, nodes may need to exchange very little information, and do so only sporadically, that is, while the number of devices in the network can be very large, only a subset wishes to transmit at any given time. Thus, a tailored protocol, which matches the signaling, physical layer and access control to traffic patterns is required.
In this work, we design and analyze a data collection protocol based on information theoretic principles. In the suggested protocol, the sink collects messages from up to K sensors simultaneously, out of a large population of sensors, without knowing in advance which sensors will transmit, and without requiring any synchronization, coordination or management overhead. In other words, neither the sink nor the other sensors need to know who are the actively transmitting sensors, and this data is decoded directly from the channel output. We provide a simple codebook construction with very simple encoding and decoding procedures. We further design a secure version of the protocol.
Information Theory