Decision Support for Operators of Fully Autonomous Systems using RESTORE: Robust Execution System for Trusted Operation in Relevant Environments

SSCI and MIT (Prof. Julie Shah) propose to develop and test a system that provides real-time assurance and trust in decisions made by autonomous collaborating vehicles. The proposed system is referred to as the RESTORE (Robust Execution System for Trusted Operation in Relevant Environments) and represents a decision support tool which facilitates decision making by the operator in cases when decisions by the Collaborative Autonomy (CA) system results in deviations from mission objectives. Key elements of RESTORE Include: (i) Contextual Interrogation Engine which uses an innovative approach to contextualize the interrogation, present the results in natural language, and increase trust in autonomy; (ii) Dynamic assurance engine which monitors mission parameters during operation and, in the case of deviations, provides the operator with alternate courses of action; and (iii) Vehicle-agnostic middleware to enable applicability of RESTORE to heterogeneous platforms operating in different domains. RESTORE will extend our field-proven CA system and MIT’s autonomy interrogation techniques with results presented in natural language. SSCI’s expertise in the development, implementation and transition of CA systems to operational platforms, combined with MIT’s expertise in machine intelligence and human-machine teaming, will result in an enhanced capability applicable to a range of DoD programs and platforms.