This book presents the underpinnings of state-of-the-art embedded systems. In doing so, it raises the bar for the next generation of ubiquitous computing. The target audience is “researchers, practitioners, and students” who are designing the Internet of Things (IoT).

Note that the book presents theories, algorithms, and formulae, not code or pseudocode. The author is interested in promoting methods and principles, not implementations. Depending on your perspective and time frame, that could be either good or bad. Reading and understanding the text is not as easy as simply downloading an open-source project, but investing effort in understanding these concepts will pay off in better products for years to come.

The text is split into ten topical chapters. Each chapter has six to 20 subdivisions, each of which presents a distinct concept backed by a formula or algorithm. The chapters are: “Introduction” (an overview of the rest of the book); “From Metrology to Digital Data”; “Uncertainty, Information, and Learning Mechanisms”; “Randomized Algorithms”; “Robustness Analysis”; “Emotional Cognitive Mechanisms for Embedded Systems”; “Performance Estimation and Probably Approximately Correct Computation”; “Intelligent Mechanisms in Embedded Systems”; “Learning in Nonstationary and Evolving Environments”; and “Fault Diagnosis Systems.” If you’re interested in any of those topics, this book provides a solid theoretical basis.

The volume itself is attractively laid out and faultlessly presented. It is definitely at the top end of current publication design and one of the best looking books I’ve seen this year. There are some occasional grammatical misconstructions, but nothing terribly distracting. Although not quite 300 pages (including references and index), it’s information-dense and reads more slowly than the average computer text. It would definitely work as a college textbook, but I’m afraid undergraduates might be overwhelmed without introductory material.