Colonies of Taskable Machines

Since each taskable machine has limited scope, achieving flexibility and sophisticated capability requires groups, or colonies, of taskable machines [2,11]. In other words, future task-solving robotic systems will be made up of colonies of taskable machines that collectively achieve a wide range of tasks. Other motivations for robot colonies include:

• Building and using several simple taskable machines can be easier, cheaper, more flexible and more fault-tolerant than building a single universal robot to accomplish all possible tasks.
• Tasks may be inherently too complex for a single robot to accomplish, or performance benefits can be gained from using multiple robots. A population of cooperating taskable machines is less limited by spatial and temporal constraints: certain problem classes are provably beyond the capability of a monolithic robot, essentially because a single robot cannot be spatially distributed (that is, in two places at the same time).
• Colonies of taskable robots are of independent research interest since their development may shed light on fields spanning the social sciences (organization theory, economics), life sciences (theoretical biology, animal ethology) and cognitive sciences (psychology, learning, artificial intelligence).

Historically, implementation of robotic colonies may have been viewed as strictly more difficult than implementation of a single robot. However, restricting robotic colonies to be composed of simpler application-specific robots can compensate for the added complexity of coordination. [16] gives example application domains. In recent years, works such as [31,35] have established frameworks for task decomposition and control in multiple-robot systems. Several of these have been implemented using real robots (see [13] for a detailed survey).