ARCORP - an Adaptive Risk-based Convoy Route Planning decision tool

This proposal describes ARCORP, an Adaptive Risk-based Convoy Route Planning decision tool that extends our Course of Action (COA) development framework and Risk-based Route Planning (RBRP) application prototype. RBRP is a Web application that enables users to plan missions, minimizing risks of attack. Using past incident data derived from SIGACT and MTF reports, the route planning application learns the risk associated with various features and fuses this knowledge with geospatial knowledge layers into the route planning process. In this effort, we will pursue two areas of significant new research that is not addressed on our RBRP project. First, we will develop trafficability models, algorithms, and a decision tool that analyzes a fused social, cultural, geospatial, and event knowledgebase. For example, ARCORP will apply knowledge from past experience learned from historical patterns of dismounted and motorized traffic to predict trouble spots associated with scheduled events such as outdoor bazaars, religious services, job recruitment queues, and other events. Second, we will develop adaptive learning mechanisms to represent dynamic events such as traffic conditions, route conditions, and attacks in our models. To support these capabilities, we will also perform research in geospatial and socio-cultural data sources, collection methods, and knowledge engineering.