Lecture 26: Robotics

Lecture 26: Robotics¶

AIMA Chapter 26 — 1 hour¶

Learning Objectives¶

• Understand robot hardware: sensors, actuators

• Apply localization and mapping (SLAM)

• Implement motion planning

• Handle human-robot interaction

Robot Hardware¶

• Sensors: Cameras, lidar, IMU, tactile

• Actuators: Motors, grippers

• Types: Manipulators, mobile, humanoid

Localization¶

• State: Robot pose (position, orientation)

• Filtering: Update belief with motion and sensing

• Particle filter: For non-Gaussian

Mapping and SLAM¶

• Mapping: Build map given poses

• SLAM: Simultaneous localization and mapping

• Loop closure: Recognize revisited places

Motion Planning¶

• Configuration space: Robot as point

• Obstacles: C-space obstacles

• RRT: Rapidly-exploring random trees

• Trajectory optimization: Smooth path

Control¶

• Open-loop: Execute plan

• Closed-loop: Feedback (PID)

• Optimal control: LQR, MPC

Human-Robot Interaction¶

• Preference learning: What does human want?

• Imitation learning: Learn from demonstrations

• Safety: Physical, cognitive

Summary¶

• Sensors, actuators

• Localization, SLAM

• Motion planning: RRT, optimization

• HRI: Imitation, preferences

References¶

• AIMA Ch. 26

• Russell & Norvig, AIMA 4e, Ch. 26

• Chapter PDF: chapters/chapter-26.pdf

Questions?¶

Next lecture: Philosophy, Ethics, and Safety of AI (Chapter 27)