Skip to content
RoboLearn
Not completed

Autonomous decision making

Control & autonomyIntermediate25 minExplore autonomously ↓Before this: State machines

What you’ll learn

  • The difference between remote-controlled and autonomous robots
  • The sense–plan–act cycle
  • How behaviours combine, with priorities

Remote control vs autonomy

A remote-controlled robot needs a person to make every decision. An autonomous robot makes its own decisions from its sensors. Many real robots are in between: they handle the details themselves and take high-level goals from people.

Sense, plan, act

  1. Sense: read the sensors and build a picture of the world (the robot’s model).
  2. Plan: decide what to do to reach the goal.
  3. Act: move, then sense again.

Simple robots skip the planning and react directly (“too close → turn”). Advanced robots build maps and plan routes.

Behaviours with priorities

A practical way to build autonomy is to stack simple behaviours, with the most important one winning:

Priority Behaviour When
1 (highest) Avoid collision Something very close
2 Follow the wall Wall on the right
3 Wander Otherwise

This is called subsumption: higher behaviours take over (subsume) lower ones when they’re triggered.

Autonomy needs fail-safes

  • Stop if sensors give impossible readings.
  • Stop if the robot hasn’t moved when it should have (stuck).
  • Always have a physical off switch.

Finished this lesson?