A stochastic approach to sensor fusion and perception control

ABSTRACT

Our work deals with robot multisensor perception and perception resources management. We first describe the stochastic model of a n close-field sensors system gathering information about its environment, and explain how Bayesian formalism applies to such a surveillance automaton. The perception control problem is here the dynamic allocation of these sensors to the different sectors of the horizon, in order to optimize the global estimation of the state. The policy that we propose was tested on a real multisensor robot with an original hardware and software architecture. We try and demonstrate the usefulness of this approach to surveillance and target detection, so that it will eventually become part of a complex system performing various perception requests in an indoor environment.

REFERENCES

• 1: A. G. Barto , R. S. Sutton , C. J.C.H. Watkins, Learning and Sequential Decision Making, University of Massachusetts, Amherst, MA, 1989
• 2: Richard Ernest Bellman, Dynamic Programming, Dover Publications, Incorporated, 2003
• 3: J.-L. Desnoyer, O. Dessoude, A. Lanusse, O. Michel and B. Zavidovique. Mastering perception heterogeneity: a balance between control, physics and computers. In Proceedings of the Sixth Scandinavian Conference on Image A~alysis, vol. 2, pp. 874- 877, Oulu, Finland, 1989.
• 4: J.-L. Desnoyer, O. Dessoude, A. Lanusse and X. Merlo. Robot perception systems: acoustics-vision set-up control, in Proceedings of the IFAC Symposium on Advanced I~formation Processing in Automatic Control, vol. 2, pp. 229-233, Nancy, France, 1989.
• 5: J.-L. Desnoyer, O. Dessoude, A. Lanusse and B. Zavidovique. Mastering multisensor perception ~ a practical implementation of stochastic fusion. In Proceedings of the IASTED International Symposium om Robotics and Manufacturing, pp. 216-220, Santa Barbara, CA, 1989.
• 6: J. C. Gittlns. Multi-armad bandit allocation indicas. John Wiley & Sons, Chichester, UK, 1989.
• 7: P. R. Kumar , Pravin Varaiya, Stochastic systems: estimation, identification and adaptive control, Prentice-Hall, Inc., Upper Saddle River, NJ, 1986
• 8: L. V. Lindhy. Scoring rules and inevitability of probability. Intaraational Statistical Raview, vol. 50, pp. 1-26, 1982.
• 9: X. Merlo. Techrdques probabilistes d qnt~gratiog et de contr~le de la percep~io~ en rue de ~on exploitation par le systems de dgcision d'un robot. PhD Thesis, Institut National Polytechnique de Lorraine, Nancy, France, 1988.
• 10: X. Merlo, J. Dupont de Dinechin and B. Zavidovique. Mukisensor stochastic integration and control. In Proceedings of the IEEE International Conference on Control and Applications, pp. WP- 5-6, Jerusalem, Israel, 1989.
• 11: X. Merlo, A. Lanusse and B. Zavidovique. ContrSle optimM d'un syst{me de perception robotique. In Proceedings of the IASTED Symposium on Applied Controlj Filtering and Signal Processing, Geneva, Switzerland, 1987.
• 12: D. J. Spiegelhalter. Probabilistic reasoning in predictive expert systems. In Kanal and Lemmer, Eds, Uncertainty in artificial intelligence, Elsevier Science Publishers, 1986.

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.9 Robotics
          Subjects: Sensors

Additional Classification:
  D. Software
  D.4 OPERATING SYSTEMS
      D.4.7 Organization and Design
          Subjects: Distributed systems; Real-time systems and embedded systems

  F. Theory of Computation
  F.1 COMPUTATION BY ABSTRACT DEVICES
      F.1.2 Modes of Computation
          Subjects: Probabilistic computation

  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE

General Terms:
Algorithms, Design, Experimentation