The rapid evolution of micromechanical fabrication techniques and other sensor, effector, and processing technologies will soon make it possible to employ large numbers of very inexpensive autonomous mobile robots with fairly limited sensor capabilities to perform real- world missions in the ground, air, space, and underwater environments. One approach to such a system is to realize desired emergent collective group behaviors with simple sensor-based reactive planners. The initial thrust of this effort has been to develop generic ensemble behaviors, such as blanket, barrier, and sweep coverage, and various deployment and recovery modes, which can address a broad spectrum of generic applications, both military and civilian. However, while different applications may require similar group behaviors, the sensor, information, and communications resources available to the participating individual robots may be very different. This paper outlines the many-robot approach to real-world problem solving and discusses the various roles that different types of sensors can play in such systems. Analysis and simulation results are presented to show how useful behavioral algorithms can be designed to make use of diverse information resources, and the area search problem is analyzed to derive both system measures of effectiveness and system design considerations.