Robot Path Planning based on Probabilistic Foam
Path Planning, Probabilistic Foam, Bubble, Front Wave, Search Tree.
Motion planning is a basic problem approached in robotics research. The capacity of define a sequence of actions to drive a robot from an initial state to a final desired state, avoiding obstacles in the workspace, is a fundamental ability, required to the development of different autonomous systems, such robotic manipulators, wheeled robots or unmanned aerial vehicles. Each motion planning method has a specific strategy to explore the workspace and plan a path. The method proposed in this work utilizes a Probabilistic Foam for robot path planning.
In the proposed method, the free space is approximately covered by a set called Probabilistic Foam, which is composed by overlapping convex subsets, called bubbles. Starting from the initial robot configuration, the bubbles of the foam propagate randomly along the free space, such as a front wave, forming a search tree structure, in order to reach the desired final configuration.
In this way, it is possible to find relatively short paths, In this way, it is possible to find relatively short paths, for which, the robot maneuvering space can be easily quantified by means of the sizes of the bubbles crossed by them.