Efficient Distributed "Hormone" Graph Gradients

Abstract

Several researchers have attempted extract the principles behind hormone gradients found in biological system, and apply them to control physically manifested distributed systems. This paper presents an efficient implementation of a graph-based version of hormone gradient mechanisms. The algorithm is based on hop-counting of the topological distance between any given vertex and the hormone emitting vertex, and can deal with dynamic changes to the topology of the system. Performance of the described algorithm is documented through a number of experiments. The algorithm was developed for use in self-reconfigurable robotics, but might very well be useful for many other applications. The use of the algorithm to provide a common coordinate system for a collection of selfreconfigurable robot modules is described, that provides pose relative to the emitting module for all modules affected by the hormone.