Andrew Viterbi simply wanted to fill in the blanks in several theories when he developed the Viterbi Algorithm. Little did he anticipate its widespread applicability in error-correcting codes in 2 billion cell phones, magnetic recording, most satellite TV receivers, a variety of cable TV systems, voice recognition, and even analyses of DNA sequencing. "When a machine understands your voice, some aspect of my algorithm is in there, improving the accuracy," he says.
Viterbi has enjoyed science ever since grade school. His father was his first influence. "He was a scholar in Greek and Latin, a physician, and a scientist, but most of all, a great father," Viterbi says. Born Andrea Viterbi in Bergamo, Italy, his family came to the United States in 1939 to escape Italy's anti-semitic laws, and they changed his name to Andrew when he was naturalized.
Viterbi's immigrant father was glad his son was going into engineering, though he would have preferred a career in chemistry. Still, Viterbi made his parents proud. As a student at Boston Latin, he did well in his studies of the classics as well as mathematics. "I had good teachers in high school," he says modestly.
Living close to the Massachusetts Institute of Technology, he was inspired to go to college there. It was a good move. "I had worldfamous professors there who influenced me. It was the time, 1952 to 1956, when engineering was morphing into applied science and becoming more solidly based," he says.
Viterbi studied statistical communications theory-in his final year and, in a cooperative program, worked on early computers with some local industries. This led to several job offers after he got his master's degree. Among them were an offer from Hughes Research Laboratories and another from the California Institute of Technology's Jet Propulsion Lab. He chose the latter. "There were more opportunities in California, and that's where I met my wife," he says.
It was a wise decision. Four months after joining JPL, the Russians launched Sputnik, which led the U.S. to boost its investment in science and grant JPL a contract for satellite work. Viterbi was a junior member of a team that provided telemetry for Explorer 1. This involved work on the efficiency of transmission of information over very large distances and consequently in very low signal-to-noise environments. He used this work on early digital communications systems in his PhD research on error correction codes at the University of Southern California. This subsequently gave him ideas for where his algorithm could be applied.
10 years of insight—and more
Once Viterbi got his PhD in electrical engineering in 1962, he went into academia. For 10 years, he taught digital communications and information theory at UCLA's School of Engineering and Applied Science. "These were probably my most productive years. It's when I wrote the most important papers (on digital communication theory) and the algorithm," he says. He wanted to find a simple algorithm to explain processing techniques. Today, the Viterbi algorithm is widely used in error-correcting codes in cell phones, speech recognition systems, DNA research, and other applications of Hidden Markov models.
Among those papers was "Error bounds for convolutional codes and an asymptotically optimum decoding algorithm," published in IEEE Transactions on Information Theory in 1967. It was decades ahead of its time! Computers then weren't advanced enough to apply the algorithm to decoding problems, and it didn't find widespread application until digital and wireless communications hit the airwaves.
From scholar to businessman
"I got the bug to become an entrepreneur," he says, founding Linkabit Corp. with Irwin Jacobs and Leonard Kleinrock in 1968. The all-digital communications company envisioned a Ku-Band satellite network architecture allowing large businesses to establish low-cost, low-to medium-speed data communications enterprise networks linking thousands of locations across the country. Many space and microwave technology companies now manufacture these systems.
Viterbi and his colleagues left the company in 1985 to start another the same year—Qualcomm. Today, Qualcomm is a world leader in innovative digital wireless communications products and services. Core to its success is Viterbi's code division multiple access (CDMA) technology. It lets many users share the same radio frequencies and increases system capacity many times over analog system capacity, eliminating the problems of congested radio waves. It also is critical to the success of today's exploding wireless industry.
Will there be enough U.S. engineers?
Viterbi sees a future filled with easy access to information. "Information—it will be ubiquitous and uniformly available to everyone," he says. He also expects to see "literacy spreading throughout the world, and with the volume of information available, that will hopefully bring peace.
"The spread of information is already having an impact on China. It's amazing how much the Internet has penetrated there, even though it has been partially blocked," he says. "You can't have very rapid industrial development without the spread of information."
The United States, however, faces an enormous challenge to maintain its outstanding role in science and technology, he warns. "We used to be the unchallenged leader. Now many other countries in Europe and Asia are challenging us."
U.S. science and technology, "which has fueled our economy for the last 50 years, is in need of repair," he says. "We need to attract more young people into the field."
See Associated Figure
Reprints
