Polymorphic Computing Structures for Distributed Sensing

One of the key uses for Wireless Sensor Networks is event detection, and pattern recognition is one of the primary methods of accomplishing this. As information processing at a centralised point is inefficient in terms of energy usage and communication, pattern recognition should take place within the network and in a distributed manner. We are developing new algorithms for performing distributed pattern recognition and classification in WSNs, based on the Polymorphic Computer (PmC) concept [1]: a device that can reconfigure itself to match the changing computing needs of the user, by rearranging its internal components. In contrast to previous approaches to implementing a PmC, we are carrying out computations primarily through message passing; to aid in this, we use an innovative communication technique based upon the concept of hyper-dimensional binary vectors [2].

References

  1. David Hentrich, Erdal Oruklu, and Jafar Saniie. Polymorphic Computing: Definition, Trends, and a New Agent-Based Architecture. Circuits and Systems, 02(04):358-364, 2011.
  2. [2] Denis Kleyko, Nikita Lyamin, Evgeny Osipov, and Laurynas Riliskis. Dependable MAC Layer Architecture Based on Holographic Data Representation Using Hyper-Dimensional Binary Spatter Codes. In International Workshop on Multiple Access Communications, pages 134-145. Springer Berlin Heidelberg, 2012.