ARCHIMEDES: Active Reasoning Conducted with Heterogeneous Information Monitors for Evading Dangers and Extended Surveillance

Abstract

ARCHIMEDES is a suite of algorithms for planning and task allocation to monitor and surveil a geometrically defined area with a group of heterogenous agents. We focus on maximizing the information gained over a fleet of such agents. The multi-step process allows for our force-based system to find a solution faster by generating reasonable starting points for agents. We present our proposed algorithms, compare them to the most relevant related work via a walkthrough discussion, and describe experimental results from ARCHIMEDES in many mathematically abstract scenarios.

Keywords

• task allocation
• information gain
• constraint reasoning
• planning