5
$\begingroup$

Does the literature on swarm robotics define the smallest number of robots required to make a "swarm".

Is it 10, 20 or 100 robots, or does the name depend on something other than the number of robots?

$\endgroup$
  • 2
    $\begingroup$ A single robot couldn't be a 'swarm', but I can't see why a pair of robots couldn't exhibit most of the traits of swarm behaviour. $\endgroup$ – Mark Booth Sep 20 '17 at 9:50
7
$\begingroup$

A robot swarm is defined by its behaviour, rather than the number of robots it contains. Swarm robots:

"... are coordinated in a distributed and decentralised way"

The key feature of swarm behaviour is self-organisation, i.e. the emergence of a global, complex pattern of behaviour based on local-level interactions between low-level, "simple" but autonomous components of the system.


"Swarms" are typically described in the literature as "large" numbers of robots. However, there is no real reason why much smaller groups of robots couldn't exhibit "swarm behaviour".

In the case of just two robots, they would probably be described as "paired robots", rather than as a swarm. However, I have seen groups of three robots described as "swarms" in research papers (for example, in this 2015 description of a Counter Abstraction Technique for the Verification of Robot Swarms from UCL).


It should be noted that in many applications, a distinction is drawn between "multi-robot systems", where inter-robot communication is not essential to the operation of the system, and true "robot swarms", which are defined by inter-robot communication. The two systems may involve similar, large numbers of robots, but are distinguished on the basis of their behaviour. This idea is explored in some depth in this 2011 paper on multi-robot exploration.

$\endgroup$
  • 1
    $\begingroup$ I agree with everything except the comment about inter-robot communication. Swarm behaviors often have only implicit communications from a small neighborhood, while multi-robot systems typically have explicit communications from most (if not all) of the team. $\endgroup$ – combo Sep 20 '17 at 15:51
  • $\begingroup$ @combo That may be true in some cases, but in my experience so far, the literature (as with the paper quoted) generally seems to describe "multi-robot systems" in terms of explicit communication with a central "controlling intelligence", and "swarms" in terms of inter-robot communication (although not necessarily communication between all of the robots in the swarm). $\endgroup$ – sempaiscuba Sep 20 '17 at 18:55
  • 2
    $\begingroup$ The presence of a central node distinguishes "centralized" and "decentralized" systems, but multi-robot versus swarms is more of statement about the "complexity" of the robots. Multi-robot systems often allow for very detailed communications (e.g. complete maps in exploration problems) versus simple implicit communication used for formation control $\endgroup$ – combo Sep 20 '17 at 19:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.