PROACTV: Projective Plan Recognition with Orbital Analysis of Conjunctive Threats and Vulnerabilities

National security requires maintaining awareness of the space environment, a task that currently relies on radar and photometric data to track the thousands of active satellites and debris in orbit. Despite the volumes of data and accurate orbit tracks produced by this continual effort, it does not provide enough contextual information to fully characterize and understand important space events. A maneuvering red satellite, for example, may be adjusting its orbit as part of a benign mission, or to intercept one of our blue satellites. To provide effective space situational awareness (SSA), human analysts require information and tools to discern emerging attacks from a sea of data with speed and confidence. To address this challenging problem for the USSF and USAF, the SIFT, UMN, Raytheon/BBN team proposes to continue our successful development of PROACTV– Projective Plan Recognition with Orbital Analysis of Conjunctive Threats and Vulnerabilities.  The PROACTV design integrates the recent success of three SIFT technologies – Project7, ELEXIR, and SAVANT – and incorporates novel UMN insights on constrained estimation. In Phase 1 PROACTV developed an innovative proof of concept demonstration in data fusion and decision support for space event characterization, combining novel SAVANT algorithms for automated spacecraft reachability analysis and maneuver prediction with our ELEXIR projective plan recognition framework.  We demonstrated fusion of data from both “hard” astrometric data sources and “soft” sources such as news feeds, web pages, and notional intelligence intercepts. Our results were significant enough to attract Raytheon/BBN to join our team for Phase II. Raytheon plans to extend their existing SSA products with PROACTV technology.   SAVANT provides an orbital analysis toolset that finds intercept opportunities for red satellites, using both traditional Lambert-based Delta-V maps and relative orbit maneuver planning software to compute both when and how they could maneuver to reach blue satellites. ELEXIR provides a massively-scalable plan recognition system to infer plans from their actions. Project7 provides the human-machine intelligence interface to address information overload and support interactive data analysis.  SIFT and BBN have already integrated Project7 into the next generation of intelligence analyst’s tools for DIA. PROACTV Phase I showed human-machine teams rapidly distilling large amounts of hard orbital data and diverse sources of soft contextual data to generate explanations of red satellite behavior and proactively mitigate potential threats. PROACTV assesses co-orbital threats by considering: intent, capability and opportunity. PROACTV combines multi-INT data sources to estimate red satellite capability, searches the space catalog and computes intercept opportunities for all red satellites based on the inferred capability, and generates and ranks alternative hypotheses that explain observed red satellite behavior.