Membrane Systems for Robot Control Architectures

Membrane systems are computational models inspired in basic characteristics of biological membranes concerning the way they process chemical compounds, energy and information. They represent a new distributed parallel computation model where the cooperation between objects (membranes) is possible to solve complex problems. Membrane systems exhibit a hierarchical topology where well-delimited computer units coexist. On the other hand, there are several control architectures for robots, classified in according their characteristics. A special class is the behavior based architecture, represented by the dynamic schema hierarchy model, the subsumption architecture, DAMN, etc. The idea behind this class of architectures is developing autonomous robotic behaviors by means of hybridizing perception and control. In this paper we propose the use of membrane systems as a behavior based robot architecture to exploit the parallel computing characteristic of membranes. We try to map behaviors to membranes to build the proposed architecture. The developed architecture was tested in an educational robot