Autoloader – An Augmented Reality Enabled Decision Support Tool with Genetic Algorithm Optimizer

The Autoloader is a decision support tool designed to allow loadmasters to efficiently create tie-down patterns while also increasing the safety and reliability of the patterns. The tool will be built with Unity3D. There are several benefits to using Unity 3D. This will allow us to take advantage of the 3D rendering to create a display that accurately represents the cargo area of an aircraft, while using color, sizing, and rendered text to provide understandable real-time feedback about the quality of the tie-down pattern. Additionally, Unity3D provides an Augmented Reality framework we can use to allow the loadmaster to add items to the library of cargo types by using a tablet’s camera and touch screen. Finally, Unity3D supports a wide variety of platforms, allowing us to develop for Android as well as enabling commercialization efforts on other platforms, including IOS and Web Clients.   Autoloader uses a mixed-initiative Genetic Algorithm to generate tie-down patterns for a fixed configuration of cargo in a variety of aircraft. The Genetic Algorithm is inspired by neuro-evolutionary research and automated architectural design. The algorithm uses a hierarchical representation of the tie-down pattern. The evolutionary process uses separate cross-over and mutation operations applied to the high-level overall cargo configuration and the low-level pattern used for each individual cargo pattern. Reasonable initial configurations are created to increase the efficiency of the search process. The algorithm will allow the system to develop complete tie-down patterns or use a mixed-initiative approach to work with the loadmaster to complete a partially specified pattern. The system is designed to find a pattern that meets all safety constraints, while minimizing the number of tie-downs used. Additionally, the system will return the safest possible result if no fully valid solution exists.