On December 14th 2012, my PhD student Ipek Caliskanelli has received the Best Paper Award at the 3rd IEEE International Conference on Networked Embedded Systems for Every Application (NESEA 2012). The paper was a joint effort by the Computer Science and Electronics departments, resulting from a "sandpit" exercise organised by both departments to foster new research collaborations. The paper's list of authors also shows the collaborative effort behind the work: James Harbin (a research assistant funded through the sandpit exercise), myself (CS), Paul Mitchell (Elec), David Chesmore (Elec) and Fiona Polack (CS).
The main contribution of the paper was a bio-inspired algorithm to balance computation and communication workload over battery-powered distributed embedded systems. We were inspired by the hormone-based communication in beehives, and developed an algorithm which is decentralised and resource efficient. The algorithm's performance was evaluated using simulation models and a small 16-node prototype of a wireless sensor network. The paper should be soon available through IEEE Xplore, and I am happy to forward a copy to anyone who is interested to learn more. An extended version will also hopefully appear at the Hindawi Journal of Distributed Sensor Networks. Ipek is currently following up that work with a parametric analysis of the algorithm, aiming to devise a design-time methodology to tune the algorithm to simultaneously maximise service availability and network lifetime.
It is worth noticing that the initial idea for the algorithm came from a chat I had with my aunt Leocadia Falkemberg Indrusiak. She is a retired university professor in Biology (and the only Indrusiak to have named an animal species), and since I was a child she has always had the patience and the talent to explain biologic things to me, regardless of my limited knowledge on the area. This time, I came to her asking about leader selection in animal groups. Ipek and I were looking for a bio-inspired way to manage the mapping of functionality to the nodes of a wireless sensor network, and we were not very happy with our first attempt, which was based on the organisation of flight formation in migratory birds (i.e. how birds decide which one goes at the front of the "V" at each point in time). So Cadia told me about different behaviours in large mammals, spiders and social insects, but when she described the hormone-based communication in beehives I immediately knew that we could adapt it for our purposes. Upon my return to York, I discussed it with Ipek and she produced a first implementation right before the research sandpit, when the other colleagues came on-board and contributed to its development and evaluation.
Finally, I've learned that York is at the forefront in monitoring the behaviour of social insects, so maybe they will uncover some more interesting mechanisms in nature that can be useful in embedded systems as well.
