Beamlet-like data processing for accelerated path-planning using multiscale information of the environment

We consider the deterministic path-planning problem dealing with the single-pair shortest path on a given graph. We propose a multiscale version of the well known A* algorithm (m-A*), which utilizes information of the environment at distinct scales. This information is collected via a bottom-up fusion method. Comparing with existing algorithms such as Dijkstra's or A*, the use of multiscale information leads to an improvement in terms of computational complexity.