Evidential Reasoning and Model-Based Simulation for Battlefield Information Management

Providing the commander with timely, accurate, and meaningful information is a critical problem in the design and deployment of battlefield information management systems. A methodology is needed for presenting the critical information in a situationally-relevant manner, to support the commander's visualization of the battlefield and assessment of tactical alternatives. We propose a hybrid solution combining model-based simulation with evidential reasoning to provide the needed integration and abstraction of battlefield information. The proposed system consists of two principle components: An evidential/knowledge-based reasoning and prediction (ERP) module for: a) data integration, filtering, and abstraction; b) situation assessment; and c) course-of-action (COA) projection. A model-based simulation (MBS) module for maintaining a dynamic battlefield database and projecting the expected COA into the future. The ERP module's situation hypothesis set is fed to a case-based reasoning algorithm that projects the current situation into the future, using experience-based outcome prediction. This prediction drives the MBS module, which provides three key functions: 1) maintenance of battlefield information in a central database; 2) visualization of battlefield data at user-defined levels of abstraction via graphical user interface; and 3) faster-than-real-time detailed simulation of the current situation.