Catalogue of Artificial Intelligence Techniques

   

Jump to: Top | Entry | References | Comments

View Maths as: Images | MathML

Simultaneous Localization and Mapping (SLAM)

Aliases: Mapping problem, SLAM

Keywords: ekf, kalman, localization, mapping, monte carlo method, simultaneous, slam

Categories: Robotics


Author(s): Michael Eldridge

Simultaneous localization and mapping (SLAM), is one of the core problems in robotics. This is a technique that allows a robot to map its surroundings, and keep track of its current location within the world. If a robot is to construct a map of its surroundings, it must be able to do this without knowing where it is in the world. This can prove to be very challenging, but matters can be made even worse when the environment is constantly changing around the robot.

In order to map effectively, the robot must be able to take in data from a variety of sensors, such as laser range finders, sonar, and cameras; this will then translate into a map of its surroundings. SLAM makes use of particle filters, also known as Monte Carlo methods, and extended Kalman Filters (EKF) to interpret information from the sensors. EKF is currently one of the most widely used methods for implementing SLAM.

In the instance of using EKF to implement SLAM, it is usually combined with a model that can sense the landscaping and features of the surrounding environment. With this model, it is required that all of the landmarks the model will reference are distinguishable, in order to map effectively.

SLAM has not been perfected as of yet, and is still an active area of research. However, it is currently being implemented in Unmanned Aerial Vehicles (UAVís) and domestic robots.


References:


Comments:

Add Comment

No comments.