LEO – Nebulous

LEO: cLuster Evolved Observatory

LEO is an AI-driven framework for designing antenna arrays and observatory clusters. By co-optimizing element topology and relative placement simultaneously, LEO leverages evolutionary, multi-objective optimization to navigate complex trade spaces, ensuring robust performance under real-world mission constraints.

A distributed antenna array observatory deployed on a snowy Antarctic ice sheet with mountains in the background.

(Concept: Distributed observatory array deployed in extreme environments)

Goal & Motivation

LEO applies AI-guided design to distributed observatories, antenna arrays, sensing systems, and communication architectures, where performance depends not only on individual element design, but on configuration, coupling effects, environmental conditions, and mission- or application-driven tradeoffs. Evolutionary, multi-objective methods can explore these large design spaces, identify robust solutions, and quantify performance and uncertainty in ways that are difficult to achieve through expert-driven iteration alone.

A central driver is operational realism: advanced system design must satisfy aggressive performance goals within tight cost, schedule, risk, and deployment constraints. LEO supports faster, more defensible trade studies early in formulation -- when configuration choices have the highest leverage and late-stage changes are most costly, while maintaining physics-based evaluation and traceability to system requirements.

Primary objective

Enable end-to-end, requirement-driven optimization of observatory clusters by co-designing array element topology and relative placement.

Why evolutionary optimization

Array design spaces are high-dimensional and strongly coupled; evolutionary search supports rapid exploration, multi-objective trade-offs, and discovery of non-intuitive configurations with interpretable trade spaces.