Model Normative and Behavioral Normative Robots

(Information Science Master) Lecture

• (The following content mainly focuses on research in the 1990s)
• Model-based
  • Define a model for problem-solving, throw the input perceived by sensors into the model, and execute the output.
  • It’s like making the rational part of a human into a robot.
  • Example: Garbage collection robot: Perceive the position of the garbage with sensors, move towards it, and collect it.
  • The mechanism is easy to understand, but accurately matching the input/output of the model with the real world is very difficult.
    • Example: Motion equations of a robot arm: The calculations become more complicated as the number of joints increases.
  • During design:
    • Incorporate human thinking into the model.
• Action-based
  • Key point: There is no model of the environment or the robot.
  • Even with simple (primitive) mechanisms, they can appear intelligent in complex environments.
  • It’s like making the reflexive/animalistic part of a human into a robot.
  • Example: Garbage collection robot:
    • The sensor and movement are directly linked in a simple mechanism.
      • Like “move forward if it’s bright.”
      • Such reflex-like mechanisms are combined to enable garbage collection.
    • No clear functions like “identify the position of the garbage” or “move towards it” are created.
  • It becomes highly robust.
  • During design:
    • Emergence - Intelligence is emergent as a result of the interaction between the body (robot) and the world.
      • “Emergence”: accidental, fluid, etc.
    • It is a different philosophy from incorporating human intentions into the robot.
• These two types existed in the 1990s.
  • With technological advancements, there is a move towards the integration of Robots and artificial intelligence.

Robots §