RatSLAM is a bio-inspired simultaneous localisation and mapping (SLAM) system. Based on continous attractor network dynamics, RatSLAM is capable of mapping by closing loops to correct odometry error.
The original RatSLAM algorithm was designed and implemented on Pioneer robots by Michael Milford and Gordon Wyeth (see RatSLAM: a hippocampal model for simultaneous localization and mapping).
There is an openRatSLAM paper available in Autonomous Robots. This paper describes how openRatSLAM works in technical detail. If you use the code we would appreciate cites please.
The C++ RatSLAM implementation is currently being used to power the iRat robot when it is online and in recent ports of the Lingodroids project to use the iRat.
There are now two versions of RatSLAM available, both based on the same code:
We provide three datasets for the ROS version of openRatSLAM
The code is released under the GNU GPL V3. Please contact David Ball if you require a more permissive license.