Innovations in machine learning, artificial intelligence, and other supporting technologies have significantly advanced the state-of-the-art in terms of the capabilities provided by autonomous unmanned aircraft systems. However, autonomous systems pose special challenges for live tests on DoD test ranges, in that the methods and infrastructure used to verify requirements for traditional military systems are inadequate to accurately measure the performance and behavioral characteristics of aircraft controlled by an on-board autonomy engine. The Safe Testing of Autonomy in Complex Interactive Environments (TACE) system provides an on-board “Watchdog� processor that automatically detects unsafe behaviors stemming from an autonomy-generated instruction or from user-defined platform/test range limitations during live tests. If unsafe behavior occurs, then the Watchdog unilaterally assumes control of the autopilot as needed to mitigate or remediate the sensed hazard. TACE also provides a ground station for test monitoring and control and a sophisticated live-virtual-constructive (LVC) infrastructure for stimulating autonomous system behaviors of interest. This paper focuses on Strategic Capabilities Office (SCO) and Test Resource Management Center (TRMC)sponsored improvements of TACE capabilities and transition of TACE to “early adopter� test ranges. More specifically, this paper will 1) discuss the integration of TACE with the NAVAIR GUARDIAN system (resulting in a TRL 6 ground-based sense-and-avoid system for collision avoidance and airspace de-confliction), 2) discuss the SCO-led integration of TACE with AFRL’s Unmanned Systems Autonomy Services (UxAS) (resulting in a hardware agnostic test tool for ranges to safely evaluate a wide variety of autonomous systems from across the DoD on low-cost platforms), and 3) discuss ongoing efforts to transition TACE to the Edwards Air Force Base (EAFB) test range.