A Cognitive Sensor Network based on Binary Spatter Codes
Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Traditionally, packets had been the minimum unity of information that had been used on digital communications to send and receive data between two entities. In them, information is stored using a local representation: a specificc binary value placed inside a specific part of the packet. Some basic protocols like TCP/IP for network communications or 3G for radio mobile communications are packet based. However, in areas like Wireless Sensor Networks (WSNs) were complex environments with limited resources are very common, a packet network may not satisfy the needs the project in development requires. Extremely noise environments can produce a great number of packet losses making really dicult to establish a communication channel between nodes.A new way to represent information was proposed by Pentti Kanerva where the information is encoded using distributed representations. Unlike local representations,information is not tightly defined by a specific value in a specific position if not the meaning is spread along a whole high dimension binary vector. The application of distributed representation inside WSNs for communication purposes started at the Dependable Communication and Computing Group inside the Department of Computer Science, Electrical and Space Engineering at Lulea University of Technology. The initial research developed a Medium Access Protocol (MAC) using distributed data representations as messages between nodes. The results obtained through simulations showed the viability of the protocol to be used on real networks inside noisy and harsh environments.This thesis continue with the previous research done providing an implementation of an application inside WSN using distributed data representations and a learning algorithm to control the behaviour of the network. The long-term goal of this and future research is to build a zero conguration fault detection system based on distributed representations using binary spatter codes.
Place, publisher, year, edition, pages
2013. , 92 p.
Technology, Wireless Sensor Networks, Binary Spatter Codes, TinyOS, NS3, Symphony, HTML5, Websockets
IdentifiersURN: urn:nbn:se:ltu:diva-42655Local ID: 0a2da03f-e956-4e19-8075-9af8c5bb2a2dOAI: oai:DiVA.org:ltu-42655DiVA: diva2:1015878
Subject / course
Student thesis, at least 30 credits
Computer Science and Engineering, master's level
Validerat; 20131201 (global_studentproject_submitter)2016-10-042016-10-04Bibliographically approved