KNOWLEDGE-BASED SELF-RECONFIGURATION AND SELF-AWARE DEMONSTRATION FOR MODULAR SATELLITE ASSEMBLY

In this research, we aim to develop a knowledge-based self-reconfiguring laboratory demonstrator made up of cubic modules for validating self-reconfiguration strategies for modular satellites. Modular satellite technologies have the potential to reduce the costs and complexity of on-orbit assembly and servicing operations, as well as reducing waste for end-of-life satellites and required mass to orbit. Modules are sized to be 10cm cubes comparable to the 1U CubeSat form factor, and can connect and communicate at will to form larger structures. Each module can carry different payloads and perform different functions, while retaining common abilities to execute software and store information. Laboratory testing of self-reconfiguration strategies for a modular structure focuses on a key challenge essential to reconfigurable satellites: the ability to autonomously determine and achieve appropriate configurations of structures in orbit without human input. We apply knowledge-based inference methods to determining a suitable configuration of modules for a modular satellite and reasoning methods for planning the sequence of actions required to implement this configuration. In addition to satellites and space structures, self-assembling hardware represents a platform that could potentially be used to build many kinds of terrestrial robots and autonomous systems.