BN+BN: Behavior network with bayesian network for intelligent agent

Kyung Joong Kim, Sung Bae Cho

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)


In the philosophy of behavior-based robotics, design of complex behavior needs the interaction of basic behaviors that are easily implemented. Action selection mechanism selects the most appropriate behavior among them to achieve goals of robot. Usually, robot might have one or more goals that conflict each other and needs a mechanism to coordinate them. Bayesian network represents the dependencies among variables with directed acyclic graph and infers posterior probability using prior knowledge. This paper proposes a method to improve behavior network, action selection mechanism that uses the graph of behaviors, goals and sensors with activation spreading, using goal inference mechanism of Bayesian network learned automatically. Experimental results on Khepera mobile robot show that the proposed method can generate more appropriate behaviors.

Original languageEnglish
Title of host publicationAI 2003
Subtitle of host publicationAdvances in Artificial Intelligence - 16th Australian Conference on AI, Proceedings
EditorsTamas D. Gedeon, Lance Chun Che Fung, Tamas D. Gedeon
PublisherSpringer Verlag
Number of pages13
ISBN (Print)9783540206460
Publication statusPublished - 2003
Event16th Australian Conference on Artificial Intelligence, AI 2003 - Perth, Australia
Duration: 2003 Dec 32003 Dec 5

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Other16th Australian Conference on Artificial Intelligence, AI 2003

Bibliographical note

Publisher Copyright:
© Springer-Verlag Berlin Heidelberg 2003.

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)


Dive into the research topics of 'BN+BN: Behavior network with bayesian network for intelligent agent'. Together they form a unique fingerprint.

Cite this