Making Risk-Informed System Architecture Decisions in Early Autonomous System Design [video]

Abstract

There is a need for rapid design, manufacture, and deployment of autonomous systems for use in dangerous and evolving environments. DoD and its contractors are coming under increased pressure to field new or redesigned autonomous systems that are under budget and on schedule while being held to the expectation of systems needing to be reliable, robust, resilient safe, and survivable. A gap exists in conceptual system modeling methods to adequately, realistically, and cost-effectively model autonomous systems that can be used for early system architecture decisions.

This talk presents ongoing efforts in: 1) providing functional modeling methods with refined function failure analysis tools that predict failure pathways, consequences, and likelihoods; 2) developing a tool that determines sustainability of functional models based on manufacturing processes, and with ongoing work in connecting functional models with manufacturing process selection for system components; 3) analyzing autonomous system designs for vulnerability to irrational behaviors of other systems within a system of systems; 4) making prognostics and health management system architecture decisions based upon early design modeling and simulations; and 5) analyzing autonomous and paired crewed/autonomous system mission goals using manufacturing and design knowledge, risk information, and risk attitudes of operators through a functional modeling approach.

The goal of this research is to provide a toolbox of systems modeling and analysis methods for practitioners to make risk and failure-informed system architecture decisions rapidly and accurately, and develop system models in the earliest stage of design that require less redesign and result in quicker time to market, safer products, and more value to the customer.