Loop Closure in SLAM Systems

Tags • 6th Mar, 16

This brief blog post will be a quick introduction to the work I've been doing for my final year Masters project at the University of York, and what the next steps are leading up to the ever impending deadline. 

SLAM

Simultaneous Localisation and Mapping (SLAM) is one of the most discussed problems in AI and mobile robotics. The goal is to create a map of an environment, and localise an agent within that map, without previous knowledge of the area, or assumptions about the robots movement. This has huge implications for building robots that are truly autonomous, the kind of robots you've been seeing in sci-fi films for decades. Those that can assess an environment, and learn more about it as they navigate, exactly as humans do. SLAM is currently implemented in Google's self self driving cars, though at a short term memory-like capacity.

This idea took a...

Read More