In this video, we explore the fascinating world of artificial intelligence by diving into the creation of an AI-powered robotic vacuum cleaner. We'll break down key concepts like agents, sensors, and actuators, showing how to design a cleaning robot that navigates a room, avoiding obstacles and cleaning efficiently. You'll learn about agent-based AI approaches, performance measures, and the algorithms behind pathfinding.

Huge thanks to ‪@marekvalasek7251‬, for his exceptional teaching skills and invaluable collaboration on this video. His ability to simplify complex concepts made the explanations clear and the video approachable for everyone. If you speak Czech, be sure to check out his version of the video once it becomes public!

If you want to learn more about AI, I recommend checking out the book "Artificial Intelligence: A Modern Approach", which served as a reference while we were writing the script for this video.

0:00 Intro
0:51 Roadmap
1:40 Agent based approach
3:09 Agent program
4:37 Reflexive agent
5:34 What do we want from the robot?
7:05 Measuring cleaning time
8:14 Performance measure
8:34 Designing performance measure
9:47 Using performance measure
10:28 Optimizing performance measure
11:31 Mathematical models
12:12 Describing the room
12:51 Describing position of the robot
13:18 Describing orientation
15:32 Describing movement
16:04 Describing rotation
16:32 Describing sensors
17:18 Moving in a room
17:57 Discretization
18:32 What do we want from the vertices?
19:12 Circle packing
20:17 Creating vertex grid
20:53 Removing vertices outside bounding rectangle
21:30 Deciding whether point lies inside a polygon
23:17 Removing remaining vertices
23:49 Removing points near the walls
24:47 Which vertices to connect?
25:35 Neighboring vertices
26:50 Edge collisions
28:05 Segment to segment distance
28:52 Removing edge collisions
29:19 What is a graph?
30:52 Path in a graph
31:19 Path length
32:20 Shortest path
32:37 Finding shortest path
33:56 Moving along the path
34:26 Move instruction
35:14 Rotate instruction
36:36 Converting path to instructions
38:04 Finding cleaning path
39:48 The result
40:13 Evaluating the result
40:41 What can we improve?