Researchers led by Francisco Martín Rico have released EasyNav, an open-source ROS 2 framework for robot navigation, according to a paper on arXiv. Most navigation software ties mapping, localisation, planning and control to particular data formats, which makes it hard to extend, compare or reason about timing, the authors write.
In EasyNav, the components interact only through one shared navigation state, so different maps, sensors and planners can be swapped under explicit rules for compatibility and timing. The team tested it on real robots, indoors with laser sensors and on large indoor-outdoor routes over several levels with GPS and 3D lidar.
Compared with Nav2 under the same conditions, EasyNav cut reaction latency from 377 to 33 milliseconds indoors and from 71 to 8 milliseconds outdoors, while using less processing power and memory.



