The fundamental behaviors of the Human Model allow the agent to handle the exceptional conditions and trained activities of their mission while performing a task. Our Human Model is a very different approach from the scripting methods used in many training systems. In scripted simulation systems the players must be instructed how to handle each contingency down to the lowest cognitive detail. However, the Human Model provides a rich cognitively functional player as the starting point thus enabling agents to perform high-level tasks and respond naturally to a dynamic environment.
Presented here are some key behaviors:
Moving in Complex Spaces: Individual agents are aware of their environment and know how to move in general terms throughout that environment. This includes navigating stairs in multi-story buildings, avoiding obstacles and avoiding threats.
Driving and Riding in Vehicles: In addition to walking in complex spaces, the agents also know how to drive ground, air and water vehicles. This includes groups of personnel using the same vehicle to travel to the same location without any special coordination being required when the tasks are assigned. The agents in the vehicle also know how to use the mounted weapon systems for that vehicle, and do so automatically if threatened.
Operating Weapons, Explosives and Equipment: Expanding upon the movement knowledge, individuals also know how to use weapons, explosives and equipment to perform tasks or to protect themselves. They can select from the weapons in their personal inventory the most appropriate choice for the given set of targets they have deemed threatening. When making these decisions, agents consider all weapons-exclusion zones and make engagement decisions based upon initial or commander updated rules of engagement.
Working Collaboratively With Other Players: Individual team members performing a set of common tasks work collaboratively to perform those tasks. For example if a group of forces must breach a door in order to enter and clear a building, those forces complete that task working together. The team leader selects a person to perform the breach based upon available tools and skills, that agent performs the breach either completing the task or becoming killed or wounded. If the breach is not completed, another agent is selected to work on the task. While the breach is being performed the remainder of the team provides cover and supporting fire.
Picking Up and Dropping Equipment as Needed: A subtle and powerful aspect of collaboration is how equipment and other carried resources are maintained within the group. Unlike scripted-tools where equipment must be provided to all personnel, if an individual carrying a required piece of equipment is killed or wounded another individual on the team diverts to pick up that equipment. Conversely, when a piece of equipment is no longer needed for the mission the agents abandons that equipment.
Self Preservation: As individuals become dynamically aware of threats in their environment through sensory updates or communication they act to protect themselves and preserve their own life. As they approach a threat or become aware of it within their proximity they move into cover and concealment locations, and actively seek to engage the threat to either suppress or neutralize that threat. In addition, as teams move under threat they work together and move in formations or using a bounding-over watch technique to advance.
Communicating Status and Intelligence Information: As an agent successfully senses and detects a threat or other important conditions they communicate that information to their direct commander, and as needed to their peers. In turn the direct commander relays information up the command chain to their commander or to subordinates. As such the flow of information is represented explicitly within the exercise. The simulation can represent limits of communication systems, active jamming or protocols that have an effect on intelligence reporting and situational awareness during the battle.
These automated behaviors associated with the Human Model provide the basis for a robust problem solving agent. Because tasks can approach the behavioral representation at a higher level they become more straightforward to implement.
Currently the task library has over thirty tasks to choose from and a player can be assigned any number of tasks in sequence. Adding additional tasks to the library is not complex and training classes are available for developers. The task library includes tasks to support movement, signaling, waiting for conditions, breaching, patrolling, defending a location, convoy operations, crowding/milling about, using mortars/RPGs, placing IEDs, detonating IEDs, fire control, cordon, sniping, using smoke obscurants, and cyber attacks among others. This library continues to grow and represents opportunities for collaboration.