Robotic Behavior

We use agent-based methods to realistically simulate the process by which robots assess and react to their environment and learn to achieve operational objectives. Behavior algorithms are evolved within the Sandia high-performance massively parallel computing environment. We use an object-oriented genetic programming (GP) engine that is integrated with Sandia’s Umbra modeling and simulation framework. These integrated tools can be used for a wide variety of simulation tasks (e.g., aerial reconnaissance, region mapping, and target search) that are needed to develop, analyze, test, and control systems and networks of intelligent machines. Contact: Mark Boslough, 505-845-8851; mbboslo@sandia.gov