Customer Reviews

Embedded Control Systems in C/C++

5 star:		(1)
4 star:		(1)
3 star:		(0)
2 star:		(0)
1 star:		(2)

 

 

This book is a complete joke, the title is a complete misrepresentation of the book's contents. I've read journal articles that have more information on embedded control than this 200+ page book.

Everything in this book can be gotten from more complete references on control systems, such as Dorf and Bishop's "Modern Control Systems" for starters. Even Phillips and Nagle's "Digital Control System Analysis and Design" has more information on embedded control when it isn't even the primary focus of their book.

Instead of explaining how to program control systems in C/C++ the author provides a brief overview of control techniques and MatLab code.

The author never actually discusses embedded control which is the title of the book. The point of the book, and even the title, is supposedly to explain how to control embedded systems with C/C++, yet the author never discusses anything remotely close to this, the closest he ever gets are programs that convert MatLab files to C. This is completely useless if the author doesn't explain how to use a converted MatLab file on an embedded microcontroller. The author doesn't even have examples on choosing processors for embedded systems.

Save $70 and put it towards a real controls book.

The author undertakes to teach you control systems w/o underlying mathematics. Well, I thought, that's nice (and, needless to say, it makes a good selling point), but... there's one but here: it may not be possible. And indeed the author has not succeeded at his stated goal. Not only that, he hasn't even tried: without any preliminaries (other than the declaration of intent to make it simple) the author starts bashing you over the head with s-notation (Laplace), poles and zeros, state space, swiped sines, linear/non-linear DEs, root locuses, eigenvalues, covariant matrices. etc. If using complicated things (after promising not to) w/o explaining or even defining them first in any way means making it simple, then this book is a success. I however, don't see it this way.

Here's the missing reality check: in order to understand what's going on there you need to be comfortable with: complex, vector, and matrix analyses, integral transforms (Laplace and Fourier), and signal processing. In other words, you need to have studied control (and that, of course, implies a solid mathematical background). You also need to know Matlab, or at least understand how one works with it in general.

The C++ aspect of the book is a joke: there's two and a half primitive snippets of code that any idiot could write if he knew what is to be achieved -- and for that you need the things I've mentioned above. Why it's in the book I can't fathom... maybe to have a pretext, no matter how thin, for putting the "in C/C++" on the cover.

Well, OK, the first few chapters (like, three) do present a few useful things (PID controllers and tuning) in an understandable way (how's your calculus, btw?). The bibliography is decent too: small, but pointed and of good quality (though no surprises here: for integral transforms, go for Churchill; for system identifications, Ljung; for signals, Oppenheim... but we already knew that, didn't we). In fact, why stop here: there are very good control textbooks as well -- Astrom, Ogata, etc., etc. Full of math, yes, but not promising anything their authors have no intention of fulfilling. Once you're through the texts like that, you'll be able to digest Mr Ledin's book (of course the question then becomes, what the hell you'll need if for?)

Another question: what's the deal with the word "embedded" in the title? There's zilch in this book about anything specifically embedded. Another sales-enhancing magic word?

Bottom line: being smart enough to understand this book implies that you don't need it; not being smart enough... well, then it seems like you don't need it either. I have no doubts that Mr Ledin is a very smart engineer -- and god bless, more power to him here -- but as to bookwriting... What good is this book? Who's it for? No simple mortal fits the bill: everyone will be either too smart to need it, or too stupid to understand it.

Regrettably, "Embedded Control Systems in C/C++" is yet another typical CMP book -- expensive, misrepresented, and useless. Oh yeah, and it's poorly written too (but that's the least of this book's flaws).

Absolutely not recommended.

I'm still finishing the book, so this review may be a little unfair. There are two unfortunate things about this book. The title indicates C/C++. There's not that much code in the book, especially for the analysis of the control systems. More implementation of the controller, which is OK. Keep in mind that's not the main goal of the book. The overview of control systems is pretty good with just enough of the theory and math. I was seeking an introduction so I could better understand what the control systems experts are battling with servo development and tuning. The second issue I have is I think the dependency on Matlab is a little too much. It would have been great if the publisher could have included a stripped down version of Matlab on the CD. Many of the 'solutions' are pretty magical as we just plug into Matlab and have our answer with somewhat incomplete explanations of what Matlab is doing to figure it out. I have yet to see an example of motor control and/or the plant model for a motor, which is what I'm particularly interested in. An example, with hardware, would have been great. Overall, I'm happy I bought the book and feel I better understand the basic theory and challenges of control system development and modeling.

I am extremely pleased with this book. It takes the reader on a path from PID controllers to more advanced topics without burying them in the details. In fact, so far, ALL of the Matlab/Simulink examples actually work and make sense.

I especially liked the Chapter 6.1 intro where he starts with PID, progresses through classical control system design and pole placement, and then proceeds to Optimal Control and Kalman Filters.

For someone who works with implementation, this means "More Code" as you progress, which means faster controllers (or DSPs) with more and more memory.