Multilevel Darwinist Brain in Robots - Initial Implementation

In this paper we present a Cognitive Mechanism called MDB (Multilevel Darwinist Brain) based on Darwinist theories and its initial application to autonomous learning by robotic systems. The mechanism has been designed to permit an agent to adapt to its environment and motivations in an autonomous way. The general structure of the MDB is particularized into a two level architecture: reasoning and interaction. This structure corresponds to a generic cognitive model where world, internal and satisfaction models are used to select strategies that fulfil the motivation of the agent. The main idea behind the proposal is that all of the components of the mechanism are obtained and modified through interaction with the environment in real time by means of on line Darwinist processes, allowing for a natural learning curve. The mechanism is able to provide solutions based on experience or original solutions to new situations. The knowledge used by the agent is acquired automatically and not imposed by the designer. Here we discuss the basic operation of the mechanism and demonstrate it through a real example in which a hexapod robot is taught to walk efficiently and to reach an objective in its surroundings.

[1]  Francesco Mondada,et al.  Evolution of homing navigation in a real mobile robot , 1996, IEEE Trans. Syst. Man Cybern. Part B.

[2]  G. Edelman.  Neural Darwinism: The Theory Of Neuronal Group Selection , 1989 .

[3]  Rodney A. Brooks,et al.  A Robust Layered Control Syste For A Mobile Robot , 2022 .

[4]  Jean-Pierre Changeux,et al.  Learning by Selection , 1984 .

[5]  Richard J. Duro,et al.  Applying synaptic delays for virtual sensing and actuation in mobile robots , 2000, Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium.

[6]  M. Conrad,et al.  Evolutionary learning circuits. , 1974, Journal of theoretical biology.

[7]  David J. Israel,et al.  Plans and resource‐bounded practical reasoning , 1988, Comput. Intell..

[8]  David Chapman,et al.  Planning for Conjunctive Goals , 1987, Artif. Intell..

[9]  Leslie Pack Kaelbling,et al.  An Architecture for Intelligent Reactive Systems , 1987 .

[10]  M Conrad,et al.  Complementary molecular models of learning and memory. , 1976, Bio Systems.

[11]  X. Yao,et al.  How to Make Best Use of Evolutionary Learning , 1996 .

[12]  J. Changeux,et al.  A theory of the epigenesis of neuronal networks by selective stabilization of synapses. , 1973, Proceedings of the National Academy of Sciences of the United States of America.

[13]  David Chapman,et al.  Pengi: An Implementation of a Theory of Activity , 1987, AAAI.

[14]  Randall D. Beer,et al.  Biologically inspired approaches to robotics: what can we learn from insects? , 1997, CACM.

[15]  J. Changeux,et al.  Selective stabilisation of developing synapses as a mechanism for the specification of neuronal networks , 1976, Nature.

[16]  Pattie Maes,et al.  The agent network architecture (ANA) , 1991, SGAR.