Collision-free path planning in the Belousov-Zhabotinsky medium assisted by a cellular automaton

Abstract.

We offer a new approach to computing a shortest collision-free path in a space containing obstacles, using an experimental chemical processor, based on the Belousov-Zhabotinsky (BZ) reaction. The chemical processor was then coupled via optical links with a two-dimensional cellular automaton (CA) processor. In the BZ chemical processor obstacles are represented by sites of local stimulation generated by an array of silver wires. Circular excitation waves are generated which travel through the medium and approximate a scalar distance-to-obstacle field. The field is taken as the initial configuration of the CA processor, which calculates a tree of 'many-sources-one-destination' shortest paths using wave spreading in a discrete excitable medium. We describe a hybrid (experimental chemical and software based) parallel processor (with parallel inputs and outputs) which uses the principles of wave-based computing in both the physical and computational levels of its architecture.