This book is very coherent in its exposition of ideas and reads almost like an "authored" book. There are some redundancy in explanation of ideas by different authors, but proper references are made to other chapters in the book (that were written by other authors) for a complete explanation.



You can find a self contained explanation of Extended Kalman Filter, Unscented Kalman Filter, and Particle Filter as applied to machine learning, where you have some parameter values to be automatically identified such as in weights for neural networks.



My interest was primarily in Unscented Kalman Filter and the book was detailed enough so that I could code my own Unscented Kalman Filter and reproduce some examples in the book. In the process, I had to look up on the internet on Robbins-Monro Algorithm because the book lacked a detailed explanation about it even though it was a suggested method for updating innovation covariance. Overall, the explanations were clear, and it has been a smooth process from reading this book to applying the algorithms to my own problem at hand.