Adaptively Informed Trees (AIT*)

Source Code

ABIT* is now in the Open Motion Planning Library (OMPL) and will be part of future binary releases. You can find details on how to download and install OMPL on their website.

Description

Adaptively Informed Trees (AIT*) is an almost-surely asymptotically optimal path planning algorithm. It uses an asymmetric bidirectional search in which both searches continuously inform each other. This allows it to find initial solutions quickly by allocating computational resources on promising paths. AIT* outperforms other almost-surely asymptotically optimal algorithms (e.g., RRT* and BIT*) on problems with expensive edge evaluations by finding initial solutions as fast as RRT-Connect and converging to the optimum in an anytime manner.

Publications

  1. M. P. Strub, J. D. Gammell. “Adaptively Informed Trees (AIT*): Fast asymptotically optimal path planning through adaptive heuristics.” in Proceedings of the IEEE international conference on robotics and automation (ICRA), 31 May – 31 Aug. 2020.