A Grid-based Approach for Finding Conditional Shortest Paths in an Unknown Environment

The problem of nding a shortest path through a known environment has been well studied. A more realistic and challenging problem is navigation through an unknown environment, where path planning is implemented \on the y" as new obstacle information becomes available through external sensors. A conditional shortest path is a path of shortest Euclidean distance that is computed based on the available environmental information at a given time. This paper solves this problem and proposes an eecient method for propagating a circular artiicial path planning wave to assign \eeective cost" information to various locations in an unknown environment. The proposed method removes many of the limitations previously associated with grid based path planning approaches, converges in real-time, and does not place any unrealistic constraints on obstacles or on the environment.