A Morphogenetic Approach to Self-Reconfigurable Modular Robots using a Hybrid Hierarchical Gene Regulatory Network

In this paper, we present a morphogenetic approach to selfreconfiguration of a lattice-based simulated modular robot, CrossCube, under dynamic environments. A hybrid hierarchical controller inspired by the embryonic development of multi-cellular organisms is proposed to form different patterns for modular robots to adapt to environmental changes. The first layer is a rule-based controller to generate a number of appropriate target patterns (i.e. configurations) for various environments. The second layer is a gene regulatory network (GRN) based controller to coordinate the modules of CrossCube to transform from its current pattern to the target pattern. This hybrid hierarchical control framework is distributed in the sense that each module makes its own decisions based on its local perception. The global behavior of modular robots emerges from the local interactions with the environment and between the modules. The simulation results demonstrate that the proposed system is efficient and robust in adaptively reconfiguring modular robots to adapt to the changing environment.