Introduction to SLAM

Let's start with a simple introduction to SLAM.

What is Robot Mapping?

1. Robot: a device that moves through the environment
2. Mapping: modeling the environment
3. Localization: estimating the robot's location

What is SLAM?

SLAM is the combination of mapping and localization. It is computing the robot's poses and the map of the environment at the same time. It is short for Simultaneous Localization and Mapping. Both the robot path and the map are unknown and those are correlated. It looks like a chicken or an egg problem because a map is needed for localization and pose estimation is also needed for mapping. That means SLAM is a fundamental problem for truly autonomous robots. It is the basis for most navigation systems.

SLAM Applications

SLAM is central to a range of indoor, outdoor, air and underwater applications for both manned and autonomous vehicles. For example, vacuum cleaner and lawn mower at home, surveillance with unmanned air vehicles, reef monitoring underwater, exploration of mines underground and terrain mapping for localization in the space.

Definition of the SLAM problem

The robot's controls and the observation are given and we want to know the map and the path of the robot.

It is hard to know the robot's motion and observations exactly. There is some uncertainty. It is more useful to use the probability theory to explicitly represent the uncertainty.

Taxonomy of the SLAM problem

1. Volumetric vs. Feature-Based SLAM
2. Topologic vs. Geometric maps
3. Static vs. Dynamic environment
4. Small vs. Large uncertainty
5. Single Robot vs. Multi Robots
and etc.