A geometric feature relation graph formulation for consistent sensor fusion

Abstract

A generic framework for sensory fusion is presented. A sensor-independent feature-based relational model, the geometric feature relation graph (GFRG), is developed for representing sensory information acquired by various sensors. Sensory fusion is accomplished through consistent integration of multiple irregular GFRGs acquired by various sensors into a regular GFRG. In the integration process. An effective, robust procedure for identifying coincident measurements of features based on uncertain geometric information and topological constraints, and a nonlinear programming formulation for the maintenance of consistency are presented. The identifying procedure is established by the algorithm IDENT in a knowledge-fusing mechanism using the Dempster-Shafer theory of belief function. Computer simulations verify the validity and performance of the framework.