AUTONOMOUS MISSION PLANNING FOR UAVS: A COGNITIVE APPROACH
Amber Stenger, Benjamin A. Fernando, M. Heni · 2013
Autonomous operation is becoming an increasingly important factor for UAVs. It enables a vehicle to decide on the most appropriate action under consideration of the current vehicle and environment state. We investigated the decision-making process using the cognitive agent-based architecture Soar, which uses techniques adapted from human decision-making. Based on Soar an agent was developed which enables UAVs to autonomously make decisions and interact with a dynamic environment. One or more UAV agents were then tested in a simulation environment which has been developed using agent-based modelling. By simulating a dynamic environment, the capabilities of a UAV agent can be tested under defined conditions and additionally its behaviour can be visualised. The agent’s abilities were demonstrated using a scenario consisting of a highly dynamic border-surveillance mission with multiple autonomous UAVs. We can show that the autonomous agents are able to execute the mission successfully and can react adaptively to unforeseen events. We conclude that using a cognitive architecture is a promising approach for modelling autonomous behaviour. 1.