Topic 7: Autonomy
This topic addresses the issues related to the integration of autonomous entities, human and agent-based, into organizations, processes, and systems.
Paper 005
Abstract Title: Analysis of Trust in Resource Expenditure in C2 Systems Using the State-Space Model
Point of Contact (POC): Norman K. Ma
POC Email Address: nma@mitre.org
POC Phone Number: (979)571-7777
POC Organization: MITRE
Country: US
Authors: Norman K. Ma
Abstract:
This paper provides a computational model of C2 object trust characteristic for activity analysis. We posit that this model can be valuable to understand the complex procedures of three out of the four capability acquisition tasks: Programmatic, Technical, Contract, and Costing. We also apply this model to Operation C2 activities to analyze effect of trust on C2 processes. We codify trust as a set of directional, compostable, and transferable functions that take C2 objects as input parameters and return a numerical trust value. These trust values are in turn used in C2 activities which we model using the OODA loop construct. Using the state-space model as a computational instantiation of C2 activities, we investigate various trust conditions and their effect on the efficiency and costs of C2 operations. We further describe the transference of trust conditions amongst strategic, operational, and tactical activities. Finally, we try to apply the trust model to the Social, Cognitive, Information, and Physical battle domains with the goal of gaining insights from the various battle activities that involving the trust element.
Paper 016
Abstract Title: Mission Analysis and Anomaly Detection Under Supervised Autonomy Using Reinforcement Learning
Point of Contact (POC): Gavin Taylor
POC Email Address: taylor@usna.edu
POC Phone Number: 410 293 6816
POC Organization: US Naval Academy
Country: US
Authors: Dr. Gavin Taylor, MIDN Kawika Barabin, and MIDN Kent Sayre
Paper 027
Abstract Title: Command and Control in Network-Centric Operations: Trust and Robot Autonomy
Point of Contact (POC): Kristin E. Schaefer
POC Email Address: kristin.e.schaefer2.ctr@mail.mil
POC Phone Number: 410 278 5972
POC Organization: US Army Research Laboratory
Country: US
Authors: Kristin E. Schaefer, Susan G. Hill, A. W. Evans III
Paper 062
Abstract Title: “Designing In” the Right Amount of Autonomy in UxS
Point of Contact (POC): George Galdorisi
POC Email Address: george.galdorisi@navy.mil
POC Phone Number: 619 553 2104
POC Organization: Space and Naval Warfare Systems Center Pacific
Country: USA
Authors: Captain George Galdorisi (U.S. Navy – Retired), Ms. Jenny Pels, Mr. Bryan Tollefson, Ms. Rachel Volner
Abstract: UAS have become almost ubiquitous in U.S. military operations today, but many still wonder how they will be effectively integrated into the existing force, to succeed as a force multiplier rather than just a force enabler. To these observers, the FY 2011-2036 Unmanned Systems Integrated Roadmap’s vision of “unmanned systems seamlessly operating with manned systems while gradually reducing the degree of human control and decision making required,” remains a work in progress.
We argue that advanced, effective autonomy is a cornerstone of the Integrated Roadmap’s future vision. Indeed, more than this, budget realities make autonomy an absolute necessity if unmanned systems are to be sustainably integrated into the joint force, as it is vital to reducing the rapidly spiraling costs of manning so-called “unmanned” systems. With this imperative, the Department of Defense is making tangible progress; recent technology developments in computational intelligence and multi-sensor data fusion hold the promise of making the military’s vision of autonomy a reality.
But as advanced degrees of autonomy are achieved and these systems begin to sense and adapt and allow blue forces to act within an adversary’s OODA (Observe, Orient, Decide, and Act) loop, these systems may well have to operate within our OODA loop. They may need to adapt to changing environments quickly without waiting for human operator oversight, input, and decision-making. As UxS ultimately provide their own command and control and self-synchronization and become truly autonomous, can the U.S. military profession change quickly enough to fully capitalize on their emerging capabilities?
This paper will explore the urgent need to achieve more autonomy in unmanned systems, show examples of ongoing projects designed to enhance the autonomy of unmanned systems, describe where more autonomy is most needed, and define some of the challenges that will need to be addressed as these unmanned systems begin to operate within our own OODA loops.
Paper 063
Abstract Title: Preserving the Human Element in the Age of Autonomy
Point of Contact (POC): George Galdorisi
POC Email Address: george.galdorisi@navy.mil
POC Phone Number: 619 553 2104
POC Organization: Space and Naval Warfare Systems Center Pacific
Country: USA
Authors: Mr. Bart Everett, Captain George Galdorisi (U.S. Navy – Retired), Ms. Jennifer Pels, Ms. Rachel Volner
Paper 092
Abstract Title: Monitoring the Plans of Autonomous System Teams
Point of Contact (POC): Douglas Lange
POC Email Address: doug.lange@navy.mil
POC Phone Number: 619-553-6534
POC Organization: Space and Naval Warfare Systems Center Pacific
Country: USA
Authors: Douglas S. Lange, Space and Naval Warfare Systems Center Pacific
Jason Landsborough, Space and Naval Warfare Systems Center Pacific
Jonathan Ou-Young, Space and Naval Warfare Systems Center Pacific
Abstract: One of the critical command and control (C2) activities is monitoring and advancing the force’s operational plans. As we move towards the C2 of forces that include teams of unmanned autonomous systems, the difficulty of monitoring those plans will increase as we no longer have human brains operating the units and providing a first human review of the situation. We have completed an initial proof-of-concept utilizing an autonomics framework to aid the commander in plan monitoring. We inherit from other research, plans generated from linear temporal logic (LTL) specifications. Specific actions are generated that meet the LTL specification and the autonomous system teams perform the task steps without direct human control. The autonomics framework is designed to monitor networks using probes to extract performance data, gauges to aggregate the data, a model evaluator that evaluates the performance against a formal model of the network, and strategies for manipulating the network to correct any problems found. As a plan can be defined as a network, we have shown that we can use this capability to monitor and even correct the performance of a plan. The key is the ability to automatically generate the network model, probes, and gauges from the plan and the LTL specification. We also require the generation of LTL from structured versions of natural language a step that we are also inheriting from other robotics planning research. Put together, commanders will be able to specify force requirements, generate plans and a plan monitoring aid to allow the commander to advance large complex plans for forces utilizing teams of autonomous systems.
Paper 116
Abstract Title: COMPUTATIONAL MODELING OF C2 FOR TEAMS OF AUTONOMOUS SYSTEMS AND PEOPLE
Point of Contact (POC): Mark Nissen
POC Email Address: MNissen@nps.edu
POC Phone Number: 831-320-0418
POC Organization: Naval Postgraduate School
Country: USA
Authors: Nissen, Mark - Naval Postgraduate School
David Place - US Navy Command Third Fleet
Abstract: The technological capabilities of autonomous systems (AS) continue to accelerate. Although AS are replacing people in many skilled mission domains and demanding environmental circumstances, people and machines have complementary capabilities, and integrated performance by AS and people working together can be superior to that of either AS or people working alone. We refer to this increasingly important phenomenon as Teams of Autonomous Systems and People (TASP), and we identify a plethora of open, command and control (C2) research, policy and decision making questions. Computational modeling and simulation offer unmatched yet largely unexplored potential to address C2 questions along these lines. The central problem is, this kind of C2 organization modeling and simulation capability has yet to be developed and demonstrated in the TASP domain. This is where our ongoing research project begins to make an important contribution. In this article, we motivate and introduce such TASP research, and we provide an overview of the computational environment used to model and simulate TASP C2 organizations and phenomena. We summarize in turn the research method. Key results follow, and we conclude then by summarizing our agenda for continued research along these lines.