You have probably used your math background if you have been playing with robots for awhile. However, unless you have been doing video processing, you may not have been able to take advantage of that class in probability—until now. Probalistic Robotics takes a look at various programming techniques for robotics that employ probability-based algorithms rather than the other approaches typically used in robotics. It is possible to mix techniques, such as behavior-based programming, but the book concentrates on how probability can be employed with various algorithms.
The book tries to present its information at a number of different levels with coverage of mathematical analysis and practical application. It has sufficient examples that most readers can get something useful from reading it. Still, to get the most out of this book you will need a good background in probability, even to follow the mathematical aspects. Those that do will find this book invaluable.
Probalistic Robotics starts with coverage of filters that are utilized throughout the book. It then addresses a range of areas such motion and perception. These really show off the advantages of using probability with specific algorithms.
The book moves through a wide range of robotic technology, including mapping and localization. It presents and analyzes a number of algorithms like the GraphSLAM that solves the SLAM (simultaneous localization and mapping) problem.
There are a number of chapters on planning and control. About the only area not addressed is visual object recognition, but that is a whole other book.
Probabilistic Robotics is definitely not light reading. It could easily fill two graduate-level semesters. But if you are into robotics in a big way, then this book is worth reading. There is no guarantee the probability will be the way to go for any or all aspects of robotics, but at this point it appears to be very applicable.