EVIDENTIAL REASONING FOR BUILDING ENVIRONMENT MAPS

We address the problem of building a map of the environment utilizing sensory depth information obtained from multiple viewpoints. The desir ed representation of the environment is in the form of a finite-resolu tion three-dimensional grid of voxels. Each voxel within the grid is a ssigned a binary value corresponding to its occupancy state, We presen t an approach for multi-sensory depth information assimilation based o n the Dempster-Shafer theory for evidential reasoning. This approach p rovides a mechanism to explicitly model ignorance which is desirable w hen dealing with an unknown environment. We present results obtained f rom this approach on a laboratory stereo sequence and contrast these w ith results obtained from the traditional Bayesian approach.