Bayesian multiple target tracking, 2nd edition [Book review]

Any book with the word "Bayesian" in the title is worth a look. In this case, we are rewarded by a carefully crafted volume that explains a plethora of useful algorithms for multiple target tracking. The second edition has gained two authors and lost one. The first edition (1999) did not mention particle filters, but the new edition includes good references and discussions of this important new technology. The text is very clear, accessible, and well organized, reflecting the education of the authors (50% with doctorates in math and 50% in mathematical statistics). The book includes many practical examples from sonar, radar, and other sensors. Approximately 75% of the authors have been experts at real-world sonar tracking for many decades. The authors carefully and politely explain why they use Bayesian methods rather than fuzzy logic, Dempster-Shafer theory, or imprecise probability, but this explanation is perhaps too polite for my taste. It is refreshing to read a book that does not erroneously assert that Dempster-Shafer theory or fuzzy logic is more robust and more general than Bayesian methods. The table on page 157 nicely summarizes the methods, outputs, and assumptions of the tracking methods considered (e.g., multiple hypothesis tracking [MHT], joint probabilistic data association [JPDA], and probabilistic MHT), but in the eagerly awaited third edition, one would like to see other methods included, such as the boring old nearest-neighbor data association, nearest-neighbor JPDA, single-scan assignment, multiscan assignment, the Hough transform, and various favors of retrodiction.