1. Tatiana Kichkaylo with Dave Barnhart and Lucy Hoag USC Information Sciences Institute

2. Background Funded as ISI Inter-divisional seedling Space Engineering Research Center +Artificial Intelligence = Spacecraft Portal for Integrated Design in Real time Driven by user needs  SERC is building satellites, fast, cheap, with student power ("teaching hospital" model)  SPIDR is there to help Systems Platform for Integrated Design in Real time

3. Design problem Spacecraft missions are typically “one-off” designs, requiring high degree of non- recurring engineering between even similar design states Spacecraft design is highly interdisciplinary, and follows many sets of constraints and rules dictated by environment, payload and mission sets Components and design approaches can be reused Engineering knowledge can be encoded to automate many design choices and analysis steps "Paperless" design helps directly and also provides basis for further automation Ultimately want to let individual engineers bounce ideas of a computer system, leading to faster, cheaper and better designs Blueprints Purchase orders Launch dates … Blueprints Purchase orders Launch dates … Magic Blueprints Purchase orders Launch dates …

4. Design Geometry/mass Power Thermal Launch Orbit Flight software Mission goals Vibration Component providers Export regulations Delivery schedules Manufacturing capabilities Validation and testing (software and hardware) Payload What does it take to build a spacecraft?

5. Bigger picture Collaborative design  Information sharing and clash detection Optimization of parameters  Limited number of parameters but with high fidelity Project management  Schedules and resources SPIDR  Automate roll out and propagation of dependencies  Optimize over structures given high-level specification  Very lightweight and driven by project needs

6. Idea: Use planning AI planning is about finding a sequence (DAG) of actions that achieve a goal  Methods have preconditions  Methods expand parts of the plan Design is about finding a set of recipes that convert a specification into a complete design  Recipes have applicability conditions  Recipes modify (expand) design Both methods and recipes have parameters Both plans and designs have constraints  Constraints in design are often represented as models

7. User revises requirements and metrics as needed Reusable knowledge base of engineering rules Mission requirements, orbit N best designs (component lists and system budgets) Reusable component database Optimization metrics SPIDR Reusable data Mission-specific data Automated design synthesis Design loop

8. Rule example

9. Semantic entities Query engine Variables and constraints Constraint propagation Tokens and rules Planner Semantic layer Constraint layer Planning layer Data representation Algorithms Optimization engine User interface iLands SPIDR Mission req's Constants P. battery Implementation

10. SPIDR meets real world Nanosatellite project at USC ISI  Low costs  Unusual constraints Reality alters trajectory  Initial "vision" for SPIDR  Single optimization tool with models in the loop  Actual use  Run every several days with design files (XML etc.) posted online  Got several extensions, model tweaked constantly  Useful for dealing with ripple effects and ensuring consistency  Users want more

11. Observations No way to close all loops  Extended to dynamic models (e.g., power profile)  Stuck on thermal and geometry due to  Lack of APIs  Licensing issues  Need for user interaction One shot is not enough – need Living Design  Feed back from out-of-the-loop models, version control  Side-by-side designs  Project structure and scheduling roll out  Still want the optimization loop

12. Observations (2) Rule design is not easy  Engineers understand the concept  Tools help: Eclipse plug-in  Crossing domain boundary is hard  Tweaks and overrides encourage local thinking  SPIDR Lite to study the knowledge engineering problem Other applications: Trade space studies and tools for acquisition  SPIDR-Energy for undersea surveillance (DARPA)

13. Last slide SPIDR is about rolling out engineering recipes and optimizing over high-level structures  Glue for collaborative design, optimization, and project management SPIDR is driven by real project(s)  Some ambitions got scaled down  New extensions and ideas emerged Next steps  New [engineering] domains  New applications  More users (students?) Tatiana Kichkaylo tatiana@isi.edu