Blanket (or Field) coverage: The objective is to achieve a static arrangement of elements that maximizes the detection rate of targets appearing within the coverage area.
Barrier coverage: the objective is to achieve a static arrangement of elements that minimizes the probability of undetected penetration through the barrier.
Sweep coverage: the objective is to move a number of elements across a coverage area in a manner which addresses a specified balance between maximizing the number of detections per time and minimizing the number of missed detections per area. (A coordinated sweep can be roughly equivalent to a moving barrier, but sweep coverage can also be achieved using random uncoordinated element movements; see the analysis of section 5.)
To achieve optimal coverage, the desired ensemble behavior is the maintenance of a spatial relationship between system elements which is appropriate to the mission sensors or effectors, and which adapts to specific local conditions. The basic approach is to do with robots what is actually done today with humans in applications such as aircraft carrier deck foreign object debris (FOD) disposal: "You guys line up at arm's length distance, and then all walk along together, picking up whatever you find as you go". The motion of each element is based on the motions of the other elements, more strongly on the motions of the nearest neighbors, but with some reference to the more remote elements as well. Thus, navigational coordination is achieved by using sensor inputs (each element sensing the position of its neighbors, relative to itself), vice explicit communications channels.
In addition to coverage behaviors, which represent the "bulk steady state" behaviors of the system, it is necessary to consider and to provide for various relevant spatial and temporal "boundary condition" behaviors, such as deployment, recovery, and navigation of the group as a whole. Other factors to be considered include obstacles and traversability, randomness of behavior, and the applicability of analogies from biology (herding, schooling, immune system and pheromone mechanisms) and physics (entropy, temperature, pressure, solid, liquid, gas) in analyzing the internal dynamical states of the system [GAGE92a].
The defining feature of the systems considered in this work is that the number of mobile robotic elements is large enough that the system command/control interface must "hide" the individual elements from the user/commander of the system. The elements don't have to be small or inexpensive, but economic and technical factors make it highly likely that they will be. The level of abstraction is such that collisions between elements or of individual elements with discrete obstacles are not dealt with: a collision avoidance mechanism may function transparently, or, alternatively, the elements may be considered to be small enough that collisions between them are very unlikely, robust enough that collisions don't damage them, or expendable enough that damage to or loss of elements doesn't matter at the system/mission level.
Our abstract system, then, consists of a large number of identical elements, although this stricture can, of course, be relaxed to permit multiple castes of elements, analogous to those of the social insects. Each element possesses: (a) some measure of mobility -- this may be fairly limited, so that, for example, in an underwater application the elements may be capable of regulating only their depth in the water column whle drifting with the currents; (b) some mission-capable sensor or effector, which actually performs the desired mission function; (c) optionally, some navigational sensor capability that allows each element to measure, at least crudely, its position with respect to at least its nearest neighboring elements and/or in some reference coordinate system (this sensor may be the same as the mission sensor capability listed above); (d) optionally, some communications capability, which may make use of the sensor capability; and (e) some processing capability, which implements algorithms to navigate the element so that that the ensemble of mission-capable sensors or effectors collectively accomplish the desired mission objectives (this may involve maintaining a specified positional relationship to its neighbors).
Up to Many Robot Systems
Please address all questions and comments to: robo-web@spawar.navy.mil
Last update: 1 December 1998.
Robotics home page