Neil K. Dhingra, Orbit Logic Incorporated; Cameron DeJac, Orbit Logic Incorporated; Josh Neel, Orbit Logic Incorporated; Alex Herz, Orbit Logic Incorporated; Trevor Wolf, University of Texas at Austin; Brandon Jones, University of Texas at Austin
Keywords: Sensor Resource Management, Sensor Tasking, Space Domain Awareness, Space Situational Awareness, Value of Information, Orbit Accuracy, Catalog Maintenance
Abstract:
The proliferation of space objects and the increasingly contested nature of space as a warfighting domain means that high-quality, often time-sensitive, tracks must be maintained on many objects to enable tactical operations and other missions as well as for effective Space Domain Awareness (SDA). Such elevated requirements impose a larger burden on sensor networks. However, intelligent planning can reduce the amount of sensor time required to maintain high quality tracks on space objects by scheduling sensor collects at times and with parameters to obtain specified levels of orbit accuracy, thereby reducing the burden on sensor networks and allowing them to effectively monitor more objects. In this paper, we will measure and illustrate the impact of orbit accuracy-based intelligent sensor tasking for SDA. Orbit Logic’s Heimdall SDA tasking software has been updated to support specified levels of expected space object orbit accuracy as a requirement for the generated sensor schedule. This can be leveraged to achieve high-accuracy time-sensitive tracks on high-value targets more efficiently, facilitating time-sensitive operations requiring high orbit accuracy. In addition, it can be used to intelligently reduce sensing on low-priority objects so that sufficient accuracy is maintained with fewer collects. Orbit accuracy is determined by the existing track and the parameters of the sensor collects on them, including timing, sensor quality, sensor phenomenology, viewing geometry, atmospheric conditions, other space or astronomical objects in the field of view, and others, and environmental conditions. By considering these factors when planning, sensor time can be preserved by scheduling data collection tasks that are expected to provide high VOI and omitting tasks that are expected to provide low VOI. Rather than greedily maximizing the Value of Information (VOI) for each individual collect, Heimdall’s intelligent planning optimizes the schedule to support broad mission objectives. For example, it may schedule that a lower-but-sufficient VOI collect on one object that is relatively well-tracked to enable a higher-VOI collect on another that has a worse track. For high-interest space objects which may unexpectedly maneuver and cause large changes to their orbit, orbit accuracy requirements can be combined with persistent monitoring requirements to plan for high-quality tracks and prevent loss of custody with fewer sensing resources. Specifications on orbit accuracy and other planning factors are expressed in sets of orders, tasking requirements which Heimdall performs intelligent sensor network tasking to fulfill. Each order specifies requirements on data gathered on a space object, identified by NORAD ID or a provided ephemeris. These requirements may entail observation timing within specified time windows, observation timing around orbital events, recurring observations at a given cadence, permissible sensor phenomenologies, permissible sensors, permissible viewing geometry, permissible weather during the collect, permissible viewing conditions (e.g., object apparent visual magnitude), minimum clearing radius, specified collection duration, minimum radar cross section, other permissible sensor parameters, and, now, a target orbit accuracy at given time(s). These orders can be overlapping and may each pertain to different aspects of the collected data. For example, an operator may issue an order requiring that data is collected at a given cadence and an order requiring that orbit accuracy meets a given threshold; Heimdall will plan accordingly to ensure that both requirements are met. Alternatively, instead of in strict constraints, these factors can be used in components in the configurable SDA-specific figure of merit (FOM) to reflect catalog and/or mission objectives. Heimdall creates schedules to optimize this FOM while obeying order constraints. Heimdall can run several optimization algorithms in parallel to generate different sensor schedules and choose the schedule that scores best in terms of the FOM. Heimdall’s enhanced orders, which now support requirements to achieve specified levels of orbit accuracy at specific times, is enabled by a module developed by University of Texas, Austin (UT Austin). Orbit accuracy is derived from the expected posterior space object track estimation error covariance, computed by the UT Austin module. The software computes the matrix specifying this covariance using the current track, planned future sensor collects, and other conditions based on a user-specifiable filter; the defaults are the Extended Kalman Filter (EKF) and an epoch state filter that directly maps information gained to expected accuracy at the prescribed time. Orbit accuracy is determined as a user-specifiable function of the covariance matrix can correspond to physical properties of the covariance ellipse, such as the volume or major axis length. In this paper, we will demonstrate that Heimdall’s intelligent tasking enhances sensor network efficiency and results in plans that achieve the same or better orbit accuracy with the same or fewer sensor collects. The plans and results will be created and evaluated with model sensor networks representative of government and partner assets. To show the increased efficiency, we first generate sensor network plans with a standard approach in which Heimdall is commanded to observe each of a set of objects at a given cadence. We then measure the orbit accuracy expected to be achieved for each object, and create plans with Heimdall by commanding that it achieves that level of expected orbit accuracy or better with no constraints on collect timing. Results indicating that the baseline orbit accuracy can matched or surpassed with the same or less sensor time have already been obtained in small examples; in this paper, we will present those results along with the results of broader studies to be conducted.
Date of Conference: September 27-20, 2022
Track: Astrodynamics