Dynamic Path Planning with Spiking Neural Networks

The path planning problem is relevant for all applications in which a mobil robot should autonomously navigate. Finding the shortest path in an environment that is only partialy known and changing is a difficult problem. Replaning the entire path when the robot encounters new obstacles is computational inefficient. Starting with an initial path and than modify the path locally does not guarantee to find the optimal path. In this paper we present a new path planning algorithm, the radar path planner, which is capable of planning paths in unknown, partially known, and changing environments. Furthermore we present an efficient implementation of the path planner with spiking neurons.