by Maja Mataric
As I page through Robot magazine and marvel at the advances being made at pioneering universities, I’ve often wished I could take off from my day job and study at one of them. What would it be like to sit in on “Introduction to Robotics” at CMU, MIT, Stanford, USC, Georgia Tech., Harvard, Cornell, Rice, U. of Illinois, or many other institutions in Japan or Europe? That not being possible, and since each of us is ultimately self-taught, what might I get out of reading the text-book for the intro class? When I Googled “university Intro to Robotics” I suspected I’d get scores of universities and their courses. As it turned out, there were less than a dozen. Broadening my search criteria still only revealed about fifty university courses focusing specifically on robotics. Delving deeper into the top dozen, I was able to get the course schedule, teach- ers, syllabus, and textbooks for most of them. Where to start?
The Robotics Primer was a perfect first choice. Most of the courses I found were not freshman level; they presume a couole of years of serious math and a foundation in either mechanical or computer engineering. They dive headlong into the deep end of the pool with lots of linear algebra, trigonometry, calculus, matrix math, forward and inverse kinematics, Jacobians, path planning, dynamics, haptic, and robot vision. If, like me, you have only a vague idea of what those things mean, then this book is for you. This book discusses the essential problems of robot design, control, and behavior without getting bogged down in the details (especially the math) of their solutions. This was perfect for me as a hobbyist since my philoso- phy is “you don’t have to be a chicken to make an omelette.” The Web is full of utilities and scraps of code you can borrow to do many of the calculations for you. The first order of business is to understand the big picture of what a robot is (and is not), then climb into the body and mind of a robot to perceive the world and try to function in it from the robot’s perspective. The Robotics Primer does this very well.
Each chapter defines an area of research around a problem in robotics. Relevant terminology and concepts are introduced and defined in a way that could be understood by a middle schooler but doesn’t insult the intelligence of a college freshman. Below is a teaser of what you’ll find in The Robotics Primer.
What is a robot? “A robot is an autonomous system which exists in the physical world, can sense its environment, and can act on it to achieve some goals.” Although we think we know one when we see one, robots are harder to define than you think.
Where Do Robots Come From? Interesting question. Where does the name “robot” come from? What was the first fictional robot? What was the first real robot? The answers satisfy our curiosity and the discussion of the evolution of robots, real and fictional, parallels the essential issues of robotics.
What’s in a Robot? The essential parts that make a machine a robot and how they work together: “A physical body, so it can exist and do work in the physical world. Sensors, so it can sense/perceive its environment. Effectors and actuators, so it can take actions. A controller, so it can be autonomous.” Arms, Legs, Wheels, Tracks, and What Really Drives Them. A discussion of effectors — the “hands or feet” of robot action, and the pros and cons of actuators — the mechanism’s “muscles” that power them. Active vs. passive actuation. Motors and gears. Servos and pulse width modulation. The principle of “degrees of freedom.” Holonomic vs nonholonomic.
Move it! The essential principles of loco- motion. Stability — balance and center of gravity. Gait — legs and their movement. Wheels and steering. Trajectories and paths.
Grasping at Straws. How useful is a robot without hands? How does a robot know where and how to reach, or how to
grasp without dropping or crushing? What are the mechanics of manipulation? Teleoperation. Kinematics and dynamics. Types of manipulators. Human-like arms and hands.
What’s Going On? How does a robot sense itself and the outside world? What kinds of sensors are available, from the simple to the complex? What are their abilities and limitations? What to do with the data they produce? Error and noise. Sensing alone is not perceiving — signal processing. Multiple sensors.
Switch on the Light. Simple sensors, passive and active, and the surprising information you can get from them. Switches, light sen- sort, encoders, position sensors.
Sonars, Lasers, and Cameras. Complex sensors, their pros and cons. Effect of environment on sensitivity. Lasers. Robot vision: edge detection, motion detection, stereo vision. Object recogni- tion.
Stay in Control. Staying on track — feedback error correction: Proportional, Integral, Derivative. Open loop/feed forward: just go/reach where you think it’s supposed to be.
The Building Blocks of Control. Introduction to control architec- tires — strategies for controlling robots when things get compli- cated: deliberative, reactive, hybrid, behavior-based.
What’s In Your Head? How does a robot remember things about itself and the world? Representation and mapping.
Think Hard, Act Later. Deliberative control.
Don’t Think, React! Reactive control — minimal thinking, quick acting. Resolving conflict through arbitration and fusion.
Subsumption architecture: bottom-up design and control.
Think and Act Separately, in Parallel. Hybrid control — combining the best and worst of both. Dealing with change.
Think the Way You Act. Behavior-based control — letting the world tell you what to do next. Interaction dynamics. Distributed representation of world map. Landmark detection. Compass direction.
Making Your Robot Behave. Behavior coordination — arbitration vs. fusion on a higher level.
When the Unexpected Happens. Emergent behavior — when a robot does more than you thought you programmed it to do.
Going Places. Navigation — how does a robot know where it is, where it’s going, and how to get there? Localization. Coverage. Planning. SLAM: Simultaneous Location and Mapping.
Go, Team! Group robotics — the fascinating world of team robots and swarms. Big Brother control, cooperation, or every robot for itself?
Things Keep Getting Better. Can robots really learn? Should they? Pavlov’s Aibo?
Where to Next? Robots helping people; it’s what it’s all about. Bibliography. An excellent list that digs deeper into the details.
Maja Mataric’s The Robotics Primer is an excellent starting point for a budding roboticist from middle school to college (author’s bio: http://robotics.usc.edu/~maja). It’s straightforward enough to be easi- ly understood without an engineering degree yet deep enough to cover the essential issues of robot design. Actually solving those issues is an ongoing process for the individual robot hobbyist, and for the professional and academic robotics community. Author Maja J Mataric is Professor of Computer Science and Neuroscience and Director of the Center for Robotics and Embedded Systems at the University of Southern California, where she is also Codirector of the Robotics Research Lab and Senior Associate Dean for research in the Viterbi School of Engineering.
—Dr. James Phelan
Publisher: MIT Press, Amazon.com: Paperback $22.66; Kindle $17.60.