Robotics Basics - Reactive Paradigm

Reactive control appeared as a paradigm in opposition to deliberative control 20210711201454. Reactive architectures have no planning part, little to no internal representation and minimal state. The sensors measurements are mapped to an action in the effectors, similar to reflexes, and they need to have the actions in mutually exclusive conditions (only one actions is triggered), every possible state has to be mapped with an action. However, as states become more and more complex, this becomes an impossible task, and to do the mapping a set of rules is used.

Sense —> Action

Action selection can be either a single one decided by the rules or a fusion of multiple actions together if they all satisfy the rules specified. Also, it’s very important to monitor different rules at the same time, so parallelism becomes very important in reactive architectures.

Sensing and data in reactive architectures

In reactive architectures, sensing and data representations are local to each behavior, without a central model. Each behavior can have a specific data representation for what is sensing, but again, this is local to each behavior.

Reactive architecture and behaviors

Behaviors 20211217202757 are a commonly used on reactive architectures as a way to map sensor readings into actions, although some hybrid architectures can also be defined using them.

Common characteristic’s of reactive architectures that use behaviors (and usually also on those who don’t):

  • Robots are agent operating in an ecological niche
  • Behaviors serve as the basic building block for robotic actions, and the overall behavior is emergent 20210713103850
  • Only local, behaviors specific sensing is permitted
  • Systems inherently follow good software design principles
  • Animal models of behavior are often cited as a basis for these systems or for a particular behavior

Example of architectures

Notes References

20211217202757 Robot behaviors

20210711200423 Robotics Basics - Control Paradigms

20210713103850 Robotics - Emergent behaviors

20210514183815 INDEX - Robotics

References

(Murphy 2000) (Matariâc 2007)

Matariâc, Maja J. 2007. The Robotics Primer. Intelligent Robotics and Autonomous Agents Series. Cambridge, Mass: The MIT Press.

Murphy, Robin. 2000. Introduction to AI Robotics. Intelligent Robotics and Autonomous Agents. Cambridge, Mass: MIT Press.