Introduction to AI 2

by Albert Major

Introduction & Lab Context

• The lecture is about embodied cognitive robotics—robots that use both a “brain” (software) and a “body” (hardware) to be intelligent.

What is a Robot?

• A robot is more than just a machine; it must meet key requirements:
  1. Body: Physical or virtual, must exist in the world.
  2. Sensors: To perceive the environment (e.g., cameras, microphones).
  3. Actuators: Motors or mechanisms to move or act.
  4. Controller/Brain: Computer or chip (not a human brain) to process data and make decisions.
  5. Environment: Operates in a real or virtual world, interacts with surroundings.
  6. Goal/Function: Has a purpose or task (e.g., grasping, pouring, entertaining).

This slide explains what makes a robot different from ordinary machines: to be considered a robot, a system must have a body, sensors to perceive the environment, actuators to move or act, a controller or "brain" to process information and make decisions, an environment to operate in, and a specific function or goal. The lecture emphasized that only when all these elements are present—physical form, sensing, acting, control, interaction with surroundings, and purposeful action—can something truly be called a robot, not just a machine

• Example: A robot is not just a coffee machine; it must sense, act, and be goal-oriented in its environment.

Quote from Nietzsche; This introduces the key idea of embodiment—that intelligence is not just about the brain or “wisdom,” but also about the body itself. In the lecture, the speaker uses this quote to argue that a robot (or a human) needs a body to truly be intelligent, not just a brain or software. This sets up the main theme: understanding embodied intelligence in robotics, where the body and its interaction with the environment are essential for real cognition

This slide introduces the idea of embodiment, meaning that robots (or agents) must have a real, physical body to truly interact with the world. The body is built from different materials and its specific design—called morphology—shapes what the robot can do. Examples in the lecture included robots with human-like, animal-like, or practical service morphologies, showing how different shapes and materials constrain or enable certain behaviors. The key point is that intelligence and function depend not just on having a brain or software, but on how the robot’s body is built and interacts with its environment.

Embodiment & Morphology

• Embodiment: The robot must have a real body in the physical world, not just code running somewhere.
• Morphology: The physical design/materials of the robot (e.g., hard plastic, metal, soft materials).
  • Some robots are designed to look or act like animals or humans; others use plant-like designs.
• Morphological computation: Sometimes, the body itself (shape/material) “computes” or enables behaviors without software (e.g., a dead fish or passive walker robot can still move due to body mechanics, not brain/controller).

Morphological Computation & Soft Robotics

• Morphological computation is when a robot’s body offloads some control from the brain—letting physical structure handle part of the behavior (e.g., passive walkers, soft tentacle robots).
• Soft robotics uses soft, flexible materials inspired by animals (like octopuses) to grasp or move in ways rigid robots can’t.
• Example: A soft robotic hand can feel textures and adapt shape for delicate tasks, and octopus-inspired robots can reach into awkward spaces.