Intro

Question 1

What are the reasons to navigate?

Answer 1

Reach a goal
Don’t bump into things
Know where you are

Question 2

Kinds of navigation?

Answer 2

Using maps
Using directions
Using guidance

Question 3

What problem do we solve using Localization and planning?

Answer 3

Where is a position

Question 4

What problem do we solve using Perception?

Answer 4

How to recognize that position

Question 5

What problem do we solve using Locomotion kinematic?

Answer 5

How to relocate myself

Question 6

What problem do we solve using Planning?

Answer 6

Where to go next

Question 7

What is observation or sensing?

Answer 7

Measuring aspects of environment
No interpretation
No meaning
Mechanical process

Question 8

What is perception?

Answer 8

Uses observation
Interprets data
Creates derived entities
Limited meaning

Question 9

What is cognition?

Answer 9

Uses perception
Associate patterns or entities with meaning
Link meaning to possible actions.

Question 10

What approaches exist to decide what to do?

Answer 10

Classical AI

New AI

Question 11

What is the classical AI approach?

Answer 11

Model based navigation
Complete modeling
Function based
Horizontal decomposition

Question 12

What is the new AI approach?

Answer 12

Behavior based navigation
It has no modeling
It is behavior based
Has a vertical decomposition

Question 13

Is it possible to combine classical AI and new AI?

Answer 13

Yes

Question 14

What is a world model?

Answer 14

It is a knowledge based containing information about what exists in the world, how entities are related and what properties those entities have in the world.

Question 15

What kinds of maps do we have? (Explain each)

Answer 15

• Continuous metric map: Records exact position for each object.
• Grid-based metric map: Records objects for each position.
• Topological map: Records connectivity of nodes. Each node represents a position or limited area.

Question 16

What is reactive control?

Answer 16

Navigate without maps.
Relies on sensor to move around the map.
Follows pre-programmed directions.

Question 17

What is a braitenberg vehicle?

Answer 17

It’s a robot which sensors are connected directly to the motors, which results in a control of the direction of movement is done using the readings from the sensors.

Question 18

What are some Path planning algorithms?

Answer 18

• Grid-based: Discretize configuration space into grid.
• Geometric: Based on geometric calculations.
• Potential field: Goal attracts the robot, obstacles repulse it.
• Sampling-based: Sparse connectivity graph in configuration space.