Summary
Design and development programs in aerospace and defense often feature high complexity and long design times, which translates into long sustainment times to pay off the incurred development costs. When the inevitable technical refresh effort for such designs begins, the design intent for the equipment in question can often be problematic to divine as the original engineers and personnel are no longer available (or possess the requisite memories).
The Reduction of Entropy for Probabilistic Organization (REPO) AI Exploration (AIE) effort aims to develop techniques to enable analysts to process large, unstructured, and unorganized documentation datasets that contain electrical Computer Aided Design (ECAD) schematic files, mechanical CAD (MCAD) drawings, and flowchart/block diagram data.
The goal is to implement REPO techniques into tools that reduce the technical and program risk of obsolescence management and re-engineering of legacy designs.
Specific capabilities sought by REPO include:
- Conduct initial flow chart analysis and link block diagram data to text documentation in a single cohesive structure
- Automate analysis of ECAD and MCAD to identify and export notable features in graphs and human-readable summary format for analysts
- Generate chord diagram models showing relationships between dataset entities and permit an analyst to consider relative magnitudes in comparison to alternative arcs to understand previously undiscoverable connections