A biologically-inspired approach to designing wireless sensor networks

In this paper, we contend that there are significant advantages in treating some classes of sensor networks as biological-like systems-both in structural design characteristics and in operational processes. We show how this design process leads to a sensor network system that is robust to topological changes, is scaleable and self-organising and has a number of other desirable features. The operating system kOS was designed to support the operation of distributed biologically-inspired algorithms, in order to accomplish tasks in a sensor network system. We look at the design of kOS and analyse its performance. The work presented has been implemented in an environmental monitoring project and has applications in other areas.