This project fosters research on multi-agent robotic systems for search and rescue operations as its long-term goal. Currently, the project is focused on obtaining new results on outdoors perception and navigation, both for individual and cooperative robots.
In this period, the work focused on topological navigation of the ATRV-Jr land robot comprising the construction of a topologic map supported on features extracted and selected from raw data acquired with a video camera and a laser scanner, a module of topological localization and the selection of a path in the topological map aiming at optimizing a given criteria.
Other work within the project concerned the usage of discrete event systems to model the navigation of a homogeneous population of mobile robots moving in an environment composed of a number of discrete sites. Results that relate the blocking and controllability properties of the automaton modeling the complete system with the blocking and controllability properties of the smaller dimension automata modeling each robot navigating in the environment were derived.
An agent-based software architecture that intends to close the gap between hybrid systems and software agent architectures was conceived and partially implemented. The developed concepts and tools provide support for: task design, task planning, task execution, task coordination and task analysis for a multi- robot system.