Multiple vision agents navigating a mobile robot in a real world

Robot vision often needs to execute different tasks simultaneously using different camera motions. Conventional robot vision allocates its resources, a camera and computing machinery, to the multiple vision tasks by directing its attention serially to different tasks. A framework that accomplishes these tasks by cooperation of independent agents, called multiple vision agents, with their own cameras and computing power is presented. A system with four cameras moving independently of each other has been developed. Each agent analyzes the image data and controls the eye motion, such as fixation to a moving target, fixation to a static target for vision-guided navigation, or monitoring a wide area to find obstacles. The various visual functions are assigned to the agents to accomplish the task in the dynamically changing real world