ABSTRACT
Multi-sensor data fusion involves combining information and data from multiple sources to obtain the most
accurate and specific assessment of an observed environment or situation [1-6]. Examples of data fusion
applications include monitoring the health of a complex machine [7], use of multiple sources to improve
medical diagnosis [8], automated target recognition for military applications [9], and environmental
monitoring [10]. Automated data fusion systems seek to emulate the ability of humans and animals to
improve their assessment of the external world by integrating information from multiple senses. Techniques
for automated data fusion range from signal processing and image processing techniques, to statistical based
estimation (e.g., Kalman filters), pattern recognition methods (cluster algorithms and adaptive neural
networks), voting, Bayesian and Dempster-Shafer methods, to automated reasoning techniques such as
expert systems, fuzzy logic, blackboard systems, and intelligent agents. An overview of these methods is
provided by Hall and McMullen [1] and Waltz and Llinas [2]. This chapter provides an introduction to
multi-sensor data fusion including the basic concepts, models, and processing architectures, algorithms and
techniques, and applications. A brief discussion of the state of the art of data fusion is provided at the end of
the chapter.