You are on page 1of 2

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/224729839

Introduction to robotics: Mechanics and control

Article  in  IEEE Journal on Robotics and Automation · May 1987


DOI: 10.1109/JRA.1987.1087086 · Source: IEEE Xplore

CITATIONS READS

70 6,163

1 author:

Frank L. Merat
Case Western Reserve University
68 PUBLICATIONS   640 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Statistical Electromagnetic Wave Propagation View project

All content following this page was uploaded by Frank L. Merat on 31 October 2016.

The user has requested enhancement of the downloaded file.


166 IEEE JOURNAL OF ROBOTICS AND AUTOMATION, VOL. RA-3, NO. 2, APRIL 1987

Book Reviews
Introduction to Robotics: Mechanics and Control-John J. Craig written with an excellent discussion of practical limits on control
(Reading, MA: Addison-Wesley, 1985, 400 pp.) Reviewed by algorithms; however, there areno clearrobot control examplesin this
Frank L. Merat, Department of Electrical Engineering and section, which limits its utility as a textbook.
Applied Physics, Case Institute of Technology, Case Western Chapter 10 on robot programming languages is a simple introduc-
Reserve University, Cleveland, OH 44106 USA. tion to programming languages with three six-line program exam-
ples. Thechapteris very cursoryand almost seems to have been
I have found Craig’s book to be an excellently written robotics added as an afterthought.
textbook. I have actually used it in an introductory graduate coursein Overall, the book is an excellent introduction to robotics. There is
robotics whereI was previously using RichardPaul’s Robot very little concentration on technology and, instead, Craig concen-
Manipulators. Overall, I have found the text to be outstanding. The trates on robotics. The book is well written, with the examples
material is developed in a logical manner. logically placed. For a new textbook, it is almost error-free-I found
Chapter 2 reviews basic three-dimensional frames and frame only one typographical error in the book. The homework problems
orientations (such as RPY orEuler angles) andthetransforms covered the text material well, and as an instructor, I appreciated
between frames. Although this is simple material to the mechanical Craig’s relative ranking ofthe difficulty of theproblems which
engineer, this section was simply written and electrical engineering allowed meto composehomework assignments rapidly. Unfortu-
students had no trouble with the material. nately, there is no homework solution manual to facilitate further the
Chapters 3 and 4 cover forwardand inverse kinematics. Craig uses instructor’s job. A very nice touch is the inclusion of programming
the Denavit-Hartenberg convention and his examples for assigning exercises atthe end of eachchapter.Theseprogramsstart with
frames are well illustrated. The book does not address robot work implementations of simple transformsand
end with complete
envelopes and concentrates on mechanics rather than robot specifica- simulation of a three-link planar manipulator. Theprogramming
tions. Craig seems to be averse to prismatic joints-almost all of his exercises are well organized and illustrate many of the limitations and
homework and example problems use rotary joints-which allowed advantages of robot simulation; however, there is no way an ordinary
for the instructor’s use of prismatic joints in many homework and class can be expected to perform all 24 programming exercises, and
exam problems. As with most robot textbooks, this one uses several many of the exercises rely upon the solution to previous exercises.
commercial robots as examples-the PUMA 560 and the Yasukawa Having some form of solution available fortheprogramming
Motoman L-3 (instead of the usual Stanford robot). exercises would be a definite plus to instructors.
Chapter 5 introduces the Jacobian and static forces and velocities. Craig only discusses the Newton-Euler approach to robot dy-
Craig introduces velocity “propagation” (from one rotating frame to namics. Having taught both Newton-Euler and Lagrange-Euler
another) which he later uses for dynamics. Again, all examples are approaches, I found that Craig’s Newton-Euler was very easy for the
for rotaryjoints with “equivalentformulas” given for prismatic students to grasp with well-done examples.
joints, One of my major criticisms of the books is the section on control.
My favorite chapter of the book is Chapter 6 on robot dynamics. As it says on the flyleaf, the book, ‘‘...is directly of use to those
Craig uses the iterative Newton-Euler approach which is intuitively engineers developing roboticsystems.. .,” and this approach shows in
easy for students (particularly electrical
engineers) to quickly his treatment of control systems. There is no review or discussion of
understand. This is not the case with the Lagrange-Euler approach simplePIDcontrol methods; instead, Craig uses control law
which many textbooks use. Craig does not, however, acknowledge partitioning. This is fine for a reference book, but a poor choice for a
the existence of other approachesto robot dynamics, and you will not robotics textbook.
even find the classic closed-form Lagrange-Euler analysis of the two- Robot programming languages may as well not have been covered.
link planar manipulator. The chapter gives three six-line-or-less programming examples and
Chapter 7 develops trajectorygeneration in joint and Cartesian briefly mentions several commercial programming languages.
space. Chapters 8 and 9 discuss robot control. The reader expecting a In summary, the book is an excellent textbook or reference for
classical discussion of PID controllers will be disappointed here. robot kinematics and dynamics. Craig’s chapters on kinematics and
There is no discussion of simple control nor of dc motors or any dynamics are the best written and easiest to follow that I have ever
“hardtechnology.” Craig instead uses theideaofacontrol law read.Thecontrolsection is well written but is probably not
which computes forces and which can be used to partition complex comprehensiveenough for afirstcourse in robotics. Finally, no
control problems into simpleservocontrolproblems.This is an mention whatsoever is made of robot sensors-especially computer
excellent approach, particularly for force control, if the student has a vision-which is now an integral part of many robotics courses.
goodcontrolbackground.Overall,thecontrolchapters are well

0882-4967/87/0400-0166$01.OO 0 1987 IEEE

View publication stats

You might also like