Knowledge-based signal interpretation

Signal interpretation is the process of using signal data to develop a high-level description of an environment, including the objects present, their classifications, and their locations. Signal interpretation plays a central role in surveillance systems whose task is to detect, classify, and localize specific platforms (such as airplanes and ships) on the basis of signals they emit or reflect. AT&T has been working in the area of surveillance technology for 40 years and is continuing work to improve detection performance and to meet demands for processing data from distributed sensor systems. This paper describes how artificial intelligence (AI) technology is being integrated with knowledge and algorithms from previous AT&T programs to develop new signal interpretation systems for distributed sensor systems. These new systems extend platform identification and classification capabilities, fuse information over space and time, enhance system modifiability, and extend the capabilities of the human-machine interface. Plans for exploiting the ASPEN parallel processor to meet real-time processing demands are also described.