Situation Assessment-Based Search Agent

We propose to design and prototype a search agent for smart munitions that exploits multiple information sources to aid an automatic target recognition (ATR) system's target acquisition capability. Smart munitions that rely on preprogrammed search paths (the weapon's flight path over the search area) ignore the dynamic nature of a battleground and fail to exploit important information that can be monitored in-flight but cannot be foreseen at launch. For example, unexpected weather events such as sandstorms or blizzards can greatly degrade an ATR sensor's fidelity, but weapons following preprogrammed paths cannot n assessment of the current situation by monitoring events from dynamic onboard data sources such as ATR output and flight performance parameters, and correlating them with static sources such as mission planing data and digital elevation maps. If the resulting situation calls for a new search path, a trajectory optimizer generates a new path (subject to constraints such as know friendly positions) that improves the probability of accomplishing mission objectives. When path choice depends on the enemy's course-of-action, a separate course-of-action prediction module contributes constraints based on identified avenues of mobility and knowledge of the enemy's military organization.