Paul Picciano (Aptima, Inc.), Nathan Schurr (Aptima, Inc.), Gabe Ganberg (Aptima, Inc.)
Keywords: SSA
Abstract:
The call for dynamic partnerships demanded in the US. Space Policy confronts two formidable challenges. The first is evident in the lack of the adoption of technical innovations that could substantially enhance collaboration. The second category, and perhaps a greater impediment, involves organizational and social constraints that minimize information sharing. Compounding the technical challenges, the organizational barriers to collaboration present a different problem set. There is a culture in the space domain that predisposes most stakeholders to guard their information. Most owner/operators are reluctant to share asset data, whether experiencing an anomaly or just providing status updates. This is unfortunate, because the owner/operators generally have the most accurate and timely data pertaining to their satellite. Comprehensive Space Situational Awareness (SSA) requires the marshaling of disparate mission critical elements. The mission threads reliant on SSA are complex and often require analysis from a diverse team of experts with sophisticated systems and tools that may be dispersed across multiple entities including military, commercial, and public interests. Two significant trends are likely to further perpetuate this state of affairs: 1) the space environment continues to be more congested, contested, and competitive, and 2) further pressures to increase SSA Sharing with a greater number of stakeholders throughout the world. The challenge of delivering the right information to the right people, while protecting national security and privacy interests, is in need of an innovative solution. Our approach, entitled Space Collaboration via an Agent Network (SCAN), enables proxy software agents to represent stakeholders (as individuals and organizations) to enhance collaboration among various agency producers and consumers of space information The SCAN agent network will facilitate collaboration by identifying opportunities to collaborate, as well as optimize the processes given the mission context. The agent-based approach is uniquely capable of addressing the collaboration challenges from both the technical and organizational perspectives. To achieve these objectives, we are employing a modeling approach based on a Markov decision process (MDP). MDPs are very general models for optimizing decisions under uncertainty. The model was chosen because it is able to represent the uncertainty in outcomes (such as loss of satellite communication or inability for a colleague to finish a task on time) as well as the associated values of decisions made and their resulting outcomes. The key aspect of MDPs is that it is a sequential model, not only does it represent uncertainty for single decisions and outcomes, but estimates the state of the world in the future and optimizes decisions based on these future expectations. The SCAN prototype will be used to assess modeling parameters and assumptions as well as collect user feedback.
Date of Conference: September 11-14, 2012
Track: Poster