swarm intelligence

Swarm intelligence is the study of decentralized, self-organized systems that can move quickly in a coordinated manner. In nature, swarms occur naturally and scientists have studied natural processes such as ant colonization, bird flocking and animal herding to learn how discrete biological agents work together with their environment to perform a common goal.

In robotics, swarm intelligence involves taking what scientists have learned from observing nature and applying the concepts to machines. For example, a robot swarm may be made up of tiny, identical devices -- each equipped with a sensor. When information collected by one robot agent is shared with the other devices in the group, it allows the individual devices to function as a united group. Robot swarm tends to be simple and agents are often equipped with sonar, RADAR or cameras to collect information about the environment around them. 

Types of communication between swarm agents

Individual bots or agents in a swarm can communicate in a variety of ways, including:

Point-to-point communication: Information is passed from agent to agent directly to alert the swarm about locations, obstacles or goals.

Broadcast communication: One agent in the swarm sends information to the rest of the swarm at once using means like sound, light or wireless

Environmental communication: A message is left by an agent within the environment the swarm is moving through, enabling the other members to interpret that information to impact their behavior. This is similar to the way insects leave a trail of chemicals behind to lead their counterparts to a specific location.

Examples of swarm intelligence

Swarm intelligence has many use cases. One application of swarm intelligence can be seen in the tiny, drone-like cameras used for dangerous search and rescue missions. When programmed to operate together as one unit, the robots are able to perform a highly sensitive global task such as searching for survivors in destroyed areas. Swarm intelligence is also used to simulate crowds in movies and interactive systems, such as virtual reality games.

Smart dust is a term used to describe a swarm of microelectromechanical robots (MEMS) that are small enough to stay suspended in air. Scientists hope that some day, smart dust can be used to detect information about environments on other planets. Characteristics that could be tracked include chemicals, vibrations, light patterns, temperatures, acceleration speeds and magnetic fields. 

In information technology (IT), network engineers have modeled to the idea of intent-based routing (IBN) after the way ants share information. A probabilistic routing table is used to reward and reinforce successful packet routing.

This was last updated in May 2019

Continue Reading About swarm intelligence

Dig Deeper on Machine learning platforms