Intelligent Navigation and Temporal Reasoning Evaluation Platform

The number of unmanned, autonomous vehicles in the military continues to grow. The missions these vehicles perform range from intelligence, surveillance and reconnaissance (ISR) to urban search and rescue (USAR) to explosive ordnance disposal (EOD). Each of these missions requires the unmanned vehicle to operate in time and space and under the supervision of a human. To provide a natural human interface the unmanned system must represent and reason about time and space in a fashion similar to humans (or at least in a fashion translatable to humans). Spatio-temporal representation and reasoning are topics of much on-going research. However, it is difficult to evaluate and compare the competing approaches because there are no standard testbeds or scenarios. In this project we propose an Intelligent Navigation and Temporal Reasoning Platform (INTREP) that can be used to evaluate spatio-temporal representations for autonomous agents. The INTREP system consists of three main components: 1) a testbed consisting of a simulation and databases for scripts and logged data; 2) a testbed control interface that assists an operator in configuring the INTREP system and analyzing the results; and 3) one or more autonomous agents that interact with the simulation to perform missions.