1: \begin{abstract}

2: Navigation function (NF) is widely used for motion planning; such a function is bounded, analytic, and guarantees convergence due to its Morse nature, while having a single minimum point at the target. This results in a safe path to the target. Originally, NF was developed for deterministic scenarios where the positions of the robot and the obstacles are known. Here we extend the concept of NF for static stochastic scenarios. \\

3: We assume that the robot, the obstacles and the workspace geometries are known discs, while their positions are random variables.  We define a \emph{Probability NF} (PNF) by introducing an additional \emph{permitted collision probability}, which limits the risks (to a set value) during robot motion.\\

4: We apply the Minkowski sum for the continuous case when considering the geometries with the Probability Density Functions (PDF). The PDF for collision is therefore the normalized convolution of the robot geometry, the obstacles geometries and their locations' PDFs.\\

5: We give an approximation for the permitted probability for collision. We then formulate an explicit function and prove that it is indeed a PNF. Finally, we exemplify our algorithm performances, and compare its results with a conventional NF algorithm.

6: \end{abstract}