Bayes Networks for Sonar Sensor Fusion

Wide-angle sonar mapping of the environ­ ment by mobile robot is nontrivial due to sev­ eral sources of uncertainty: dropouts due to "specular" reflections, obstacle location un­ certainty due to the wide beam, and distance measurement error. Earlier papers address the latter problems, but dropouts remain a problem in many environments. We present an approach that lifts the overoptimistic in­ dependence assumption used in earlier work, and use Bayes nets to represent the depen­ dencies between objects of the model. Ob­ jects of the model consist of readings, and of regions in which "quasi location invari­ ance" of the (possible) obstacles exists, with respect to the readings. Simulation supports the method's feasibility. The model is readily extensible to allow for prior distributions, as well as other types of sensing operations.