The central goal of this research is to develop a comprehensive methodology for the design, testing/verification, and realization of multi-level logic-based switching control systems for the supervision, coordination and control of uninhabited vehicles operating autonomously in cluttered, passively-sensed environments. This research involves a cross-disciplinary blend of ideas drawn from computer vision, automatic control, and high-level programming languages. Vision and control serve as the base disciplines for algorithm development while high level languages provide the enabling software for the testing, verification, simulation, rapid prototyping and implementation of these algorithms.

    Through collaboration with several other research programs at Yale, we are focusing on the notion of autonomous coordinated motion, using analogies from nature such as flocking birds, schooling fish, herding animals, and swarming insects. We are interested in tasks such as congregation, basic navigation, obstacle avoidance, target capture/envelopment, target fleeing, and other forms of intelligent coordinated behavior. Our shared experimental testbeds include a custom-made team of robots competing in the RoboCup 2000 robotic soccer competition, a team of Probotic Cye "Luxury Wagons", and a small number of model helicopters.

    Central to our software effort is a programming methodology called Functional Reactive Programming, or FRP. With FRP it is possible to develop software for control applications in a high-level, declarative manner, and to use the same infrastructure in a simulation or prototyping environment, including the use of FRP animation to display control signals and related data. Our plan is to use, assess, and refine FRP to support the testing, evaluation and prototyping of logic-based switching control systems specific to this research
program.