C-space compression for robots motion planning

Exact motion planning for hyper-redundant robots under complex constraints is computationally intractable. This paper does not deal with the optimization of motion planning algorithms, but rather with the simplification of the configuration space presented to the algorithms. We aim to reduce the configuration space so that the robot's embedded motion planning system will be able to store and access an otherwise immense data file. We use a n-DCT compression algorithm together with a Genetic based compression algorithm, in order to reduce the complexity of motion planning computations and reduce the need for memory. We exemplify our algorithm on a hyperredundant worm-like climbing robot with six degrees of freedom (DOF).