Abstract :
Since the birth of the subsumption architecture (SA) two decades ago, several derivations have been emerged from the original concept. The behavior-based architectures, one of the successors, have been successful to create simple behaviors in a system, resulting more complex behaviors by their interactions. The dynamic behavior network (DBN), introduced in this paper, does not specify a strict hierarchy between the behaviors, however, the stimulus levels and links determinate their actual states in a moment. The dynamic instantiation and destruction of the behaviors make the design and the development of the behaviors easier and the computational resources are utilized more efficiently. The concept of the DBN is presented in this paper, the common and new behavior patterns are examined as well as a practical application based on DBN is described.
Keywords :
artificial intelligence; robot dynamics; DBN concept; artificial intelligence; behavior destruction; behavior instantiation; dynamic behavior network; robot behavior; subsumption architecture; Actuators; Artificial intelligence; Bismuth; Computer architecture; Informatics; Machine intelligence; Robots;