Adaptive all-source data fusion system development

Abstract

The sensor management is a key element of adaptive all-source data fusion and it must operate so that full advantage is taken of the strengths of each sensor. It is accomplished by (1) coordinating the data collection processes of disparate sensors to support the overall goal (assessment of the environment: displays of tracking kinematics and targets IDs), (2) properly selecting the use of active and passive sensors (emission control), and (3) improving the response time of a sensor by cueing it with information derived from another sensor. In the adaptive all-source data fusion, the sensor management will play a role as a feedback link connecting the automated situation assessment function and the sensors. We address how the sensor management can be accomplished and where the sensor management should reside in a complex sensor system. There are 3 aspects to the question: (1) architectures, (2) scheduling techniques, and (3) decision making techniques. The article focuses on the development of decision making techniques (advanced computing techniques) for performing the sensor management. The advanced computing techniques are based on distinctively different theoretical bases, for example, statistical approach, artificial intelligence (expert system), neural networks, and fuzzy logic. These techniques are discussed in the context of the adaptive all-source data fusion software development.