This work proposes a new architecture for controlling the navigation of a mobile robot, which consists of fusing the output of distinct controllers through a decentralized information filter (DIF). The output of each controller is inputted to a local filter with a covariance associated to it. The lower this covariance is, the bigger the influence of the output of the corresponding controller on the fused output is. A fuzzy logic approach is proposed to qualitatively determine such covariances. It is also carried out an analysis of the stability of the proposed control architecture. A conjecture based on energy considerations is introduced, which assures that the overall control signal emerging from the fusion engine behaves well. In order to ensure the accomplishment of such conjecture, a supervisor is included in the proposed architecture. One out of several experiments using the proposed control architecture is also presented, in order to illustrate the system performance.
[1]
Tucker R. Balch,et al.
AuRA: principles and practice in review
,
1997,
J. Exp. Theor. Artif. Intell..
[2]
M. Sarcinelli-Filho,et al.
Mobile robot navigation based on the fusion of control signals from different controllers
,
2001,
2001 European Control Conference (ECC).
[3]
Alessandro Saffiotti,et al.
A Multivalued Logic Approach to Integrating Planning and Control
,
1995,
Artif. Intell..
[4]
Paolo Pirjanian.
Multiple objective behavior-based control
,
2000,
Robotics Auton. Syst..
[5]
Teodiano Freire Bastos Filho,et al.
A NEW MOBILE ROBOT CONTROL ARCHITECTURE VIA CONTROL OUTPUT FUSION
,
2002
.