An Intelligent Mult-Sensor Fusion Architecture for Autonomous Decision Making

Duke University, Durham NC

Investigators

Abstract

Proposal Title: An Intelligent Multi-Sensor Fusion Architecture for Autonomous PI's Name: Devendra P. Garg Institution: Duke University Proposal Number: 9911195 Decision Making Project Abstract: This Small Grant Exploratory Research (SGER) project proposes an innovative and unconventional multi-sensor fusion architecture, with applications to a variety of scientific and engineering systems that acquire, process, and integrate information gathered from multiple knowledge sources and use this information for critical decision making. The proposed architecture leads to a critical enabler for autonomous operation of such systems, primarily due to its two significant attributes. Firstly, the fused information is richer, i.e., it contains information that cannot be obtained otherwise from each individual sensor acting alone, and thereby enables the system to perceive or measure its physical environment more accurately. Secondly, the new architecture can efficiently fuse the information from multiple sensors in a self-organizing, adaptive fashion, which allows it to modify or control its own operation while interacting with its environment. This attribute eliminates the need for having, a priori, detailed and accurate information about the sensors, and of their physical environment. The new architecture can be gainfully employed in unstructured and uncertain environments. In addition to developing a sound mathematical theory, computer simulations and laboratory experiments are designed and used to validate the configuration and performance of the new architecture.