When Intelligence is mentioned in relation to CGF, usually path finding and geometry avoidance issues are the main focus of the discussion; therefore sometime, while virtual humans walk around an HMMWV simulating a demonstration, it could happen to watch them continuing repetitively their activities as robots despite the fact that somebody take control of a soldier and shoot down a women in the crowd: virtual graphics could be fine, but definitively this is not realistic; what reaction is most likely expected in this case? People running away, someone get paralyzed, others, as berserks, attack armed soldiers; many factors affect the reaction of the population, even in a so small event, such as the behavior of the shooting private, the emerging leadership over the demonstrators, the previous history on soldier/population relationships; for sure, today, we are not able to predict exactly what will be the proper reaction, therefore operative people & scientists are able to draw different alternatives and with the support of M&S experts it is possible to create stochastic models able to represent this aspect; Simulation Team developed several CGF and simulators taking care of reproducing decision processes and human factors.
This problem applies even in simple movement context: traditionally CGF are defined intelligent, if they are able to move from point A to B by applying basic path finding, avoiding obstacles, using roads and considering terrain in order to minimize time and distances; therefore in IA-CGF one major focus is to develop new movement strategies for CGF considering risks factors: so a platoon executing a movement, choose its path considering its perception and information about threats such as snipers, opposite forces, terrain context, rules of engagement (i.e. avoiding contact) as well as support from their allied; this consideration it is quite trivial for a military on the field representing a major driver and it is applied continuously from everybody driving through traffic downtown, so it could be expected that intelligent CGF should address these aspects today.
IA-CGFs developed by Simulation Team address this problem by applying multiple techniques for modeling the environments and the decision making process of the intelligent agents; for sure a critical aspect is to create algorithms for reproducing human behavior modifiers as dynamic element of the simulation; for instance the extensive adoption of fuzzy logic allowed to combine multiple feeling and different rules within the same scenario based on the unit perception, status and behavioral characteristics; so for instance when two groups get in touch it is computed the mutual attitude, not in term of friend or foe (or neutral), but considering the concurrent weighted presence of indifferent, hostile, negative, positive and friendly feelings; in addition the occurring events and the simulation evolution change continuously the attitude among units and among groups and organizations to who they belongs, taking into account even the nature of the different evolution criteria (i.e. aggressiveness is accumulated slowly and explode quickly). Based on this approach ordering a patrolling within an area will generate deterrence, but will increase even tension and stress in IA-CGF; obviously the tuning, verification and validation of the simulator requires specific procedures, therefore the accumulated experience allows to succeed in this complex task with support of operative people and subject matter experts.
In fact it is evident that dealing with humans, these simulators don't represent a crystal ball, therefore simulation based on IA-CGF is able to analyze many alternative course of actions, repeating the same scenario and considering the impact of stochastic factors; so, if it is not possible to generate a perfect prediction, it becomes possible to quantify risks, costs/benefits and consequences of different hypotheses and decisions.
IA-CGF PSYOPS reproducing a medium size town
in term of population and social network
and their reaction to
and insurgent actions