I/ITSEC 2022

Subcommittee Category: Emerging Concepts and Innovative Technologies

Abstract: Today the defense industry, the warfighter, and government leaders are all looking to modeling and simulation to solve new technology and security challenges such as a) the creation of large-scale battlespaces to test new technologies and concepts b) create the vast amounts of data required to train machine learning models c) provide new ways to inform decision makers with campaign-level simulation. Unfortunately, creating complex, multi-domain simulations is a time-consuming and expensive endeavor. At CAE, we are working to solve these issues by developing a high-level, natural language framework for rapidly generating large and complex, multi-domain battlespaces and communicating with intelligent agents within simulations. A framework will be introduced where intuitive commands can be given to a simulation to generate large formations of entities as well as issue orders and commands to the formations. A system of collaborative intelligent agents then interprets these orders as mission objectives, which are then further broken down into country, unit and platform specific tactics that can be applied to the specific scenario to get a desired outcome.

Subcommittee Category: Training

Abstract: Last year’s paper “An Emulation of a Flying Boom Operator Using a Rule-Based Expert System” (I/ITSEC 2021) presented an approach to enhance the realism of the flying boom movement of a constructive tanker during a Virtual Air Refueling (VAR) training. The main contribution of that study is to document the actions that a boom operator performs to control the flying boom attitude and extension prior to, during, and upon disconnect. Additionally, the actions of a boom operator were subsequently used to derive a rule-based expert algorithm and implement it into a simulation of a KC-135 constructive tanker.

Because the boom motion and orientation are known to change the trim of the tanker aircraft during contact, further enhancement to the boom simulation will be necessary to reproduce the dynamic interaction of the receiver aircraft, the refueling boom, and the tanker aircraft. For instance, the tanker wake and atmospheric turbulence can disturb the motion of the flying boom; therefore, additional controls of the flying boom attitude will be required to enhance the emulation of the boom operator. Another factor related to the boom movement is when the boom extends, the inertia property of the flying boom will change as well. Therefore, this factor will also affect the movement of the flying boom as the aerodynamic force and moment change.

This present paper reviews and models these dynamic effects on the boom operator actions. The emulation of the boom operator presented in last year’s paper was updated accordingly. The preliminary simulation results of this study demonstrated an enhancement of the realism of the flying boom movement, both visually and operationally, and consequently improved the simulation fidelity normally required for a Virtual Air Refueling training. We will present the simulation results of the implementation and discuss the crucial lessons learned from this effort.

Abstract: The CV-22 Osprey fills an important need for increasing the United States Air Force’s (USAF) reach into remote and contested environments, but its unique design and capability profile leads to substantial operation and maintenance costs. Moreover, such a complex platform requires highly-skilled specialists to ensure optimal support of operational tempo, but limited fleet size translates to significantly constrained resources for training maintainers. Maintainers’ opportunities for hands-on training is further complicated by varying aircraft availability, task demands, and operational tempo. To supplement traditional training opportunities, Link developed the Immersive Maintenance Guide (IMG) as an adjunct training tool that promotes knowledge and skill development among new maintainers. The IMG allows 3-Level maintainers to virtually and independently “walk” through the steps involved in a broad selection of maintenance tasks. A collaboration between the Rapid Sustainment Office (RSO) and the Air Force Research Lab (AFRL) evaluated the IMG using both objective and subjective measures to quantify the holistic benefits of the system over a five-month evaluation period in comparison to maintainers who used only the traditional training resources. A comparison of the two groups revealed substantial improvements for both Experimental and Control participants, though advances over baseline were highest among those who used the IMG. This was evident for participants’ knowledge of the task steps, such that Experimental participants’ demonstrated a 42% greater improvement over their peers, and for their self-efficacy to lead completion of the task. Moreover, these gains were significant for simple and routine tasks as well as those less commonly encountered. As a demonstration of training benefits, the current assessment revealed the IMG’s promise as a valuable tool for improved task knowledge and an increased sense of readiness for task completion, gains which in turn likely contribute to increased combat power.

Subcommittee Category: Simulation

Abstract: Digital twins have become an important means of offering actionable insight into our world. The existing and potential applications are extremely broad in scope and drive a need for rapid creation of large scale digital replications of the real world.

We organize our digital twin application around four conceptual categories or “layers”: Physical, Human, Resource and Cognitive. The Physical layer represents the world “as-built” and can include imagery, terrain and 3D models. The Human layer includes the population and their location, either through simulated pattern of life or real time geolocation feeds. The Cognitive layer encapsulates human awareness, behavior, decision making, social connections and communication. The Resource layer contains the assets available for planning and simulation.

This paper reports on the design, implementation and integration of the geospatial data used to create the Physical and Human layers of one such digital twin, the city of London. Our project made extensive use of both cloud technology and open standards from the Open Geospatial Consortium (OGC). We faced multiple challenges related to: (i) the conversion of large amounts of photogrammetry data for streaming using the OGC 3DTiles standard (ii) the sourcing and processing of sufficiently detailed data for pattern of life and power network simulations and (iii) the performance challenges associated with using OGC WMS/WFS for real-time simulation.

Viewed individually, none of the challenges we faced are unique in the geospatial and simulation communities. However, we believe the intersection of the two domains and the lessons learned from our requirement to integrate data across the four “layers” may provide insight to others planning their own urban digital twins.

Subcommittee Category: Emerging Concepts and Innovative Technologies

Abstract: A confluence of technologies has made the simulation of large-scale pattern of life possible to support real time applications. Ubiquitous cloud computing, new generation application scalability frameworks, advanced crowd modelling software, geospatial data production tools and high fidelity/high performance visualization all contribute to the capability not only to simulate city and country sized populations but also do it in real time to support visual simulation applications.

Our work focuses on simulating population mobility, specifically pedestrian movement, as they go about their daily lives but also with the ability to interject events that would cause visible changes in their behaviour. Our objective is to go beyond simple clutter with performance and fidelity requirements that includes (i) a persistent population size of 100K – 1M+ independent agents and a minimum crowd size of 50K (ii) obeyance of walking areas and avoidance of other moving entities and (iii) a simultaneous visualization of the entire population and a game quality street level visualization of any area in the city. We believe this capability can contribute to any simulation and training application that can benefit from a realistically populated city, especially one that may require human interactivity. Specific examples include emergency evacuation scenarios, large crowd protests and military operations in urban terrain.

This paper begins by reviewing the significant existing work in this field that has inspired our methodology. It then reports on the design, implementation and integration of the various components that allowed us to achieve our overall objectives. Finally, it discusses the lessons learned and also looks to the future in this rapidly advancing field.

Subcommittee Category: Human Performance Analysis and Engineering

Abstract: Social media is pervasive throughout today's society. Reports of, and reactions to, events can propagate locally, nationally, and internationally almost instantly, and can themselves trigger a cascade of reactions and events. Within this reality, we believe that it is important for any large-scale simulation of an engaged population using human behaviour modelling to incorporate social media. This paper reports on the design, implementation, and integration of the social media component within a broader simulation of Greater London with application to decision support systems, the training of decision makers, and course of action analysis. Within the social media component, live social media messages were combined with synthetically generated messages and presented in real-time, while analytics aggregating sentiment expressed in the messages were displayed to decision makers throughout the course of the simulation. The synthetic messages were generated by a trained AI model; each was related to the different emotions of happiness, sadness, fear, anger, joy, surprise and assigned to a member of the simulated population. The integration of an externally triggered event in the simulation, a power outage, resulted in a change in behaviour of the simulated population and consequently a change in the resulting tone and emotion reflected in the synthetic social media messages. Finally, initial user feedback is reported, and considerations for additional factors to influence the synthetically generated social media content based on cognitive state, demographic attributes, and extensions to the AI models are discussed.