Design paradigms for control architectures

In simple terms, a robot's objective is to fulfill a mission that can be broken down into a succession of tasks to be carried out in a known or unknown, static or dynamic environment.

Classically, a robot can be broken down into three distinct parts:

• on the one hand, a set of exteroceptive and proprioceptive sensors enabling it to gather information about its environment and its estimated state, with in some cases sensors carried by the environment itself;

• a set of actuators enabling it to act or interact with its environment;

• finally, between these two ends of the control chain, a control architecture that selects the action (behavior) to be implemented, depending on the mission objective, the current state of the robot and that of its environment....