Deep Learning, Human-Robot Interaction, and Robotics Lecture Review

Flashcards providing key definitions and vocabulary terms for Deep Learning, CNN anatomy, Vision Transformers, Human-Robot Interaction, and Robotics principles.

Deep Learning (DL)

A technology that enables robots to perceive, interpret, and interact with their environment in a human-like manner, providing greater autonomy, adaptability, and efficiency.

Automatic Feature Abstraction

A key addition of deep learning that learns useful features automatically from raw data, reducing the need for manual feature engineering.

Hierarchical Pattern Recognition

A deep learning process where early layers learn simple patterns such as edges, and deeper layers combine them into complex objects or concepts.

Convolutional Neural Network (CNN)

A deep-learning algorithm designed for grid-like data such as images that automatically learns features like edges and textures.

Convolution Layer

A CNN layer that uses learnable filters to scan the image and highlight important features.

ReLU / Activation Layer

A CNN component that adds non-linearity to the network, allowing it to learn complex patterns.

Pooling Layer

A CNN layer that reduces spatial or image resolution while preserving important information to increase model efficiency.

Fully Connected Layer

The layer in a CNN that interprets learned features and produces the final decision or prediction.

Low-Level Features

Basic features learned in the early steps of a CNN, including edges, corners, lines, and basic color contrasts.

Mid-Level Features

Intermediate features learned by a CNN, including textures, curves, contours, and repeated patterns.

High-Level Features

Complex features learned in deeper CNN steps that represent specific object parts such as wheels, faces, handles, and shapes.

Object Detection

A technology enabling robots to identify, locate, and categorize objects in their environment using deep-learning models like YOLO, SSD, and Faster R-CNN.

Vision Transformer (ViT)

A deep-learning architecture that divides an image into fixed-size patches, treats them as a sequence, uses self-attention, and learns global relationships across the image.

Human-Robot Interaction (HRI)

A multidisciplinary field concerned with understanding, designing, and evaluating robotic systems for use by or with humans.

Remote Control (HRI Level)

An interaction level where the human is in full control, giving commands that the robot follows without independent decision-making.

Shared Control (HRI Level)

An interaction level where the robot uses AI and sensors to execute tasks while working together with a human who can guide, supervise, or override.

Fully Autonomous (HRI Level)

An interaction level where the robot operates completely independently, sensing its environment and making decisions using AI without human intervention.

Cognitive Load

A human factor in HRI referring to the mental effort required by the user, which should be minimized through intuitive interfaces.

Robotics

An interdisciplinary field of science and engineering concerned with the design, construction, and use of robots to perform repetitive, complex, or hazardous tasks.

Robot

A reprogrammable, multifunctional manipulator that performs different tasks through programmed motions with appropriate intelligence.

Actuators

Robot components that produce physical movement in joints, wheels, arms, and grippers using electric, hydraulic, pneumatic, or linear mechanisms.

Effectors / End-Effectors

Robot attachments or tools that perform actual tasks, such as grabbing, welding, painting, cutting, or drilling.

Controller

The central processing unit of a robot that processes sensor data, calculates movement, sends signals to actuators, and handles safety and errors.

Humanoid Robot

A robot designed to look and behave like a human, typically possessing a head, torso, arms, hands, and legs, integrated with sensors and AI for decision-making.

Robot Processing Pipeline

The fundamental operational workflow of a robot consisting of Perception (Sensors) -> Decision-Making (AI) -> Action (Actuators).