Automated Environmental Mapping with Behavior Trees

Automated exploration, environmental measuring and mapping using robotic systems is highly relevant in post-disaster scenarios. Challenging conditions and presence of unknown hazardous substances might disturb rescuers, affecting their performance or even endanger their lives. Under these circumstances, the deployment of robotic systems can reduce the risk and workload of rescuers and improve the situational awareness of emergency services. To accomplish this, multiple algorithms need to be synchronized on the robots. Behavior Trees offer a simple and modular structure for task execution, and provide a deep insight into complex task flows at first glance. The main contribution of this work is the development and integration of methods for environmental measuring and robot exploration using Behavior Trees. These methods include robot navigation, complete coverage planning, 2D/3D obstacle mapping and localization of hazard substances. Our solution is tailored for post-dis aster scenarios and focuses on providing remote situational awareness for rescuers. Our system was successfully tested at the European Robotics Hackathon 2019, where a hazardous materials incident response operation was staged. With this implementation, we aim to simplify the system design and application of robotic systems used for autonomous operation.