As my name is not a household word outside of my field intelligent systems research, it seems appropriate to provide some background as to who I am. From an early age I have been fascinated with problems of how things work and interested in how to make things work. High School physics introduced me to how force and mass interact to produce motion and transform energy into different forms. Geometry and trigonometry taught me how the structure of the world can be described and analyzed. The Radio Amateur’s Handbook provided me with a wealth of knowledge about electric circuits and vacuum tubes. As a summer student in college, I worked at the Naval Research Laboratory (NRL) in Washington, DC, in a sounding rocket program that was later incorporated into "Project Vanguard" with the intent to produce the world’s first artificial earth satellite. My assignment was to design the antenna and feed network for that satellite. Unfortunately, the Russians launched "Sputnik" before the first US launch attempt, and the first Vanguard rocket blew up on the launch pad. As a result, the Von Braun team in the NASA Huntsville space center placed the first American satellite in orbit. However, my satellite finally made it into orbit where it will remain for about two hundred and fifty years. A spare model hung in the Smithsonian Air and Space Museum for several years.
James Albus 1970
In 1958 when NASA was formed, my Naval Research Laboratory team became the core of the Goddard Space Flight Center. Over the next fifteen years at Goddard, I invented, designed, built, tested, and analyzed data from a series of electro-optical sensors that measured the orientation of spinning satellites and moon probes relative to the sun and earth. Based on this work, I was awarded a scholarship that enabled me to pursue what eventually became my lifelong scientific goal: to understand how the brain works.
In pursuit of this goal, I graduated from the Ohio State University with a master’s degree in radio astronomy and control theory, attended the University of Southern California for background in neuroscience, obtained a PhD from University of Maryland in electrical engineering and robotics, and attended Johns Hopkins and the University of Maryland Medical School to study biophysics and neuroanatomy. I also worked at the National Institute of Health to learn how to monitor the activity in single neurons in the brain that respond to visual and tactile stimuli.
As an engineer, it is my conviction that you don’t really understand something until you can build it. From the beginning, it has been my conviction that the principal function of the brain is to generate and control behavior. So I have pursued my background in control theory and robotics as an experimental tool for understanding the workings of the brain. I started at the bottom, in the spinal cord, which is the lowest level in the control hierarchy of the brain. The circuits in the spinal cord are basically servo control computers that translate commands for force and velocity into motion of the limbs and muscles. The next step up the brain’s control hierarchy is the mid-brain, where feedback from the spinal cord is combined with inertial signals from the inner ear and visual input from the eyes to generate precise, coordinated motion in all vertebrates from fish, to birds, to humans. The cerebellum enables a bird to fly at high speed through the branches of a tree without collision, and to perch lightly on a twig. It enables squirrel to jump from branch to branch. It enables a batter to hit a pitch, an outfielder to compute where the ball will come down, enables quarterback to throw a touchdown pass, and a wide receiver football with one hand while touching his toes to the ground.
In a 1967 book, The Cerebellum as a Neuronal Machine, Eccles, Ito, and Santagothai published data about the cerebellum that was sufficiently quantitative and anatomically accurate for me and David Marr to independently formulate theories of cerebellar function that have been combined to become a classic used by researchers studying the cerebellum to this day.
I used this theory of cerebellar function as a part of my PhD thesis to control a robot arm that could be commanded to perform and learn various coordinated motions. From this work, I patented a neural network called the Cerebellar Model Articulation Controller (CMAC) that was named by Industrial Research Magazine as one of the most significant inventions of the year 1976. CMAC has been used in many adaptive control systems and is still used by neural net researchers for building adaptive controllers. CMAC has the ability to rapidly learn to compute a wide variety of control functions, and can be stacked in layers to plan and execute complex behaviors. This observation led to the development of a family of intelligent control systems for robots and automated manufacturing systems. In 1973, I left NASA and joined the National Bureau of Standards (NBS) where I teamed with Dr. Anthony Barbera to develop a family of intelligent real-time control systems (RCS). RCS was fi rst used as the basic architecture for robots and machine tools in an Automated Manufacturing Research Facility at NBS. This captured the attention of Dr. Tony Tether, who was then a program manager at the Defense Advanced Research Projects Agency (DARPA) and later became DARPA Director. He suggested that we use RCS for controlling a group of combat aircraft. Unfortunately, this was too revolutionary at that time even for DARPA, so our project was redirected to use RCS to control multiple autonomous underwater vehicles. NASA took notice and contracted us to design an RCS controller for the Space Station Telerobotic Servicer.
The US Bureau of Mines contracted us to design standards for automated underground coal mine systems. Over the years, RCS controllers have been developed for the US Postal Service to control general mail facilities and automated stamp distribution systems. Commercial controllers have been developed for machine tools used by General Motors and Boeing, and for a variety of commercial water jet and flame cutting machines. Along the way, I became interested to build really large robots, and invented a series of RoboCranes© that were cited by Construction Equipment and Popular Science magazines as among the one-hundred top products, technologies, and scientific achievements of 1992. In recognition of my contributions to manufacturing technology, I was named a "Hero of US Manufacturing" by Fortune Magazine in 1997.
James Albus in the HLPR Chair
During the 1990s, Dr. Barbera and I developed controllers for a series of Army Ground Vehicles culminating in the 4D/RCS reference model architecture for Intelligent Vehicle Systems that was adopted by the Army Future Combat System (FCS) for the Autonomous Navigation System to be used on all large FCS vehicles manned and unmanned.
Over the years, I have authored more than 180 scientific papers, journal articles, book chapters, official government studies, and popular press articles on intelligent systems, robotics, and economic implications. I published several scientific papers on computation and representation in the brain that explore the possibility of reverse engineering the human brain. I have lectured extensively throughout the world on robotics and intelligent systems, am a member of the editorial board of the Wiley Series on Intelligent Systems, and serve on the editorial boards of six prominent journals related to intelligent systems and robotics.
Since the beginning with my research for Peoples’ Capitalism in 1976, I have been deeply concerned about the economic, social, and political impact of the rapidly approaching advent of super automation. In particular, I have focused on the implications for jobs and income for the average worker. In 1976, this was largely a theoretical problem. In 2011, it has become a very current and practical issue. Modern industry simply does not need millions of new workers while the population of available workers is multiplying and the need for farm workers is declining around the globe.
In the 1976 book, I presented my vision of how the benefits of technological developments in advanced automation could become the means for liberating humankind from poverty and oppression. In this current book, A Path to a Better World, I have expanded this vision into a quantitative economic model for solving the current problems of slow economic growth and rising debt, and have suggested a foundation for a future prosperity based on widespread ownership of capital assets.