Frontiers in Robotics and AI

This paper discusses a novel approach to managing complexity in a large self-assembled system, by utilizing the self-assembling components themselves to address the complexity. A particular challenge is discussed namely the question of how to deal with elements that are assembled in different orientations from each other and a solution based on the idea of introspective circuitry is described. A methodology for using a set of cells to determine a nearby cell's orientation is given, leading to a slow (0(n)) means of orienting a 2D region of cells. A modified algorithm is then describe to allow parallel analysis of/adaption to dis-oriented cells, thus allowing re-orientation of an entire 2D region of cells with better-than-linear time performance (0(sort(n))). The significance of this work is discussed not only in terms of managing arrays of dis-oriented cells but also more importantly as an example of the usefulness of local, distributed self-configuration to create and use introspective circuitry. Cross-Disciplinary Semiconductor Research (CSR) Program award from Semiconductor Research Corporation (SRC) [G15173] This work was supported by the Cross-Disciplinary Semiconductor Research (CSR) Program award G15173 from the Semiconductor Research Corporation (SRC).