Rapid high-confidence intelligent interior mapping

This paper presents preliminary results on an intelligent mapping capability that generates high-confidence layouts of building interiors from limited exterior and interior structural observables. Such a capability could provide critical support to intelligence gathering and/or life-saving operations in the military, law enforcement, disaster response, and commercial sectors where rapid understanding of unknown interior layouts is required to save lives, maintain covertness, or minimize costs. The fundamental approach relies on an intelligent rule-based inferencing process which operates on limited structural observables. The rules are based on geo-specific design practices and building codes. The mapping capability has been demonstrated experimentally in a test-case building structure using a set of imaging sensors to scan the environment. Data was gathered at sparse locations within the building to represent a real-world rapid exploration scenario. Preliminary results show that this capability can successfully generate high-confidence interior layouts from limited structural observables. It is envisioned that this intelligent mapping capability could be integrated on unmanned ground vehicle platforms or in human-borne (soldier, SWAT, urban search-and-rescue personnel) systems.