Path planning in combined 3D grid and graph environment

In research literature and many scientific disciplines, solution to the common problem in path planning for an autonomous robot has been extensively developed. Almost all explored techniques assume the robot has complete and detailed overview of the environment he is moving in. In addition to, many methods work over the graph representation of this environment which can be very difficult to construct or obtain in the real applications. This paper introduces a hybrid technique combining graph and grid representations of an examined space and capable of planning paths in known, partially known, unknown and dynamic environment at the price of the pseudo optimality of results.