NewVectors Projects
Real-time Adversarial Intelligence Decision-making (RAID) Program
NewVectors was recently awarded a contract as part of the Department of Defense (DoD) Real-time Adversarial Intelligence and Decision-making (RAID) program. The program focuses on the challenge of anticipating enemy actions in a military urban operation. Three technology teams were selected for the program. NewVectors and Stillman Associates (under separate contracts) are developing complementary approaches to the Adversarial Reasoning Module (ARM). The ARM will provide the commander with a predictive, anticipative analysis of enemy future actions while making suggestions for friendly actions. It will also monitor the unfolding operation and continuously confirm, disconfirm and update the products of predictive analysis.
The NewVectors project is appropriately entitled Synergetic Tactical Reasoners Offering Near-term Guidance: an Adversarial Reasoning Module (STRONG ARM). A third technology team, Lockheed Martin, is developing the Deception Reasoning Module that will identify enemy’s attempts to conceal its assets and actions and to deceive the friendly forces.
NewVectors is focusing on the development of the Behavior Evolution and Evaluation (BEE) module. Soar is using NewVectors’ platform to develop the Knowledge-based Intention Projection (KIP) module and Sarnoff is developing a Statistical Anomaly Detection (SAD) module.
The BEE module, using swarming techniques that have been developed on other programs, will use evolutionary programming algorithms to dynamically evolve simple agents whose behavior can be tuned by a personality vector to simulate the actual behavior of combatants in an urban environment. The BEE agents include a rudimentary emotional model that is used to estimate the actual emotional state and its impact on the future behavior of the enemy.
One of the major challenges that the team will be addressing is the huge size of the possible future states that need to be analyzed in an urban setting (estimated to be on the order of 1,050,000). The challenge of predicting enemy actions involves reasoning about the future enemy actions in a way that combines: the enemy’s plans to achieve his objectives by effective use of his strengths and opportunities; the enemy’s perception of friendly strengths, weaknesses, and intents; the enemy’s tactics, doctrine, training, moral, cultural and other biases and preferences; the impact of terrain, environment (including non-combatant population), weather, time, and space available; the influence of personnel attrition, ammunition, and other consumable supplies, logistics, communications, sensors, and other elements of a military operation; and the complex interplay and mutual dependency of friendly and enemy actions, reactions, and counteractions that unfold during the execution of the operation.
Since this tool is designed to be used by a commander in the field, the technology must be capable of rapidly assessing a large number of alternative possible futures, identifying the most likely and the most dangerous, and providing recommendations to the commander on courses of action he may take all within a matter of minutes on a laptop-class computer. Recognizing that timeliness is a critical aspect of the technology, the program office has defined specific performance gates that must be achieved before any of the teams can progress to the next phase.
The project was started on September 1, 2004, and by the first meeting held on October 14, 2004, the NewVectors team already had a running demonstration of each of the three components, connected together and exchanging messages. On December 10, 2004, the team delivered the first code to the systems integrator (SAIC) for initial testing with the other modules of the RAID system.