The Design of Robustly Implementable Autonomous and Intelligent Machines project aims to develop fast and robust autonomy solutions using the least complex models possible, including rapidly transferring these solutions to new platforms and between platforms and domains.
To meet this goal, we are developing a Design Suite of rapid autonomy transfer tools, algorithms, and software for building same-day autonomy solutions with minimal fidelity/complexity models that are robust to performance specifications and adaptable to various platforms and domains. Technical thrusts include:
- Time Critical Autonomy Transfer
- Performance Assessment, Testing, and Evaluation, and
- Model Refinement and Autonomy Redesign.
This work is led by the University of California, Berkeley with partner Stanford University. It is part of the Defense Advanced Research Projects Agency Transfer from Imprecise and Abstract Models to Autonomous Technologies (TIAMAT) program, award number HR00112490425.