Oral-History:Arogyaswami Paulraj
About Arogyaswami Paulraj
Arogyaswami Paulraj was born in Coimbatore, India, on 14 April 1944. He earned a B.E. from Naval College Engineering, Lonavala, India in 1966, and a Ph.D. from Indian Institute of Technology, New Delhi, India in 1973.
Dr. Paulraj's development of multiple input-multiple output (MIMO) antenna technology for wireless communications has revolutionized both local area and mobile broadband communications, enabling high-speed access to multimedia services. Employing multiple antennas at both the transmit and the receive stations, the success of MIMO is its ability to provide both higher data rates and wider coverage areas. He first developed the idea of MIMO in 1992 while at Stanford University. Using the spatial multiplexing concept that exploits MIMO antennas, he demonstrated that spectral efficiency could be improved by transmitting independent data streams from each transmit antenna and then exploiting the distinct spatial signatures of each stream at the receive antennas to separate them. He was issued a patent for the MIMO concept in 1994, but he faced skepticism from industry and funding sources. However, he persisted and held annual workshops at Stanford on the technology that eventually helped interest in MIMO take hold in the late 1990s. In 1998, Paulraj founded Iospan Wireless Inc. as the first company to incorporate MIMO technology in a commercial system. The lessons learned at Iospan gave the wireless industry confidence to incorporate MIMO into emerging wireless standards. Iospan’s technology underpins today’s 4G wireless systems. In 2003, Intel Corp. acquired part of Iospan to help launch its own push into wireless broadband, further establishing the importance of Paulraj’s MIMO concept.
Paulraj is an IEEE Fellow, and in 2013 was awarded the Alexander Graham Bell Medal of the IEEE “For pioneering contributions to the application of multiantenna technology to wireless communication systems.” He is a Professor Emeritus at Stanford University, Calif. and a senior advisor to Broadcom Corp., Irvine, Calif.
About the interview
AROGYASWAMI PAULRAJ: An Interview Conducted by Michael Geselowitz, IEEE History Center, 15 November 2022.
Interview #885 for the IEEE History Center, The Institute of Electrical and Electronics Engineers, Inc.
Copyright Statement
This manuscript is being made available for research purposes only. All literary rights in the manuscript, including the right to publish, are reserved to the IEEE History Center. No part of the manuscript may be quoted for publication without the written permission of the Director of IEEE History Center.
Request for permission to quote for publication should be addressed to the IEEE History Center Oral History Program, IEEE History Center, 445 Hoes Lane, Piscataway, NJ 08854 USA or ieee-history@ieee.org. It should include identification of the specific passages to be quoted, anticipated use of the passages, and identification of the user.
It is recommended that this oral history be cited as follows:
Arogyaswami Paulraj, an oral history conducted in 2022 by Michael Geselowitz, IEEE History Center, Piscataway, NJ, USA.
Interview
INTERVIEWEE: Arogyaswami Paulraj
INTERVIEWER: Michael Geselowitz
DATE: 15 November 2022
PLACE: Virtual
Early life and education
Geselowitz:
Hello. This is Michael Geselowitz from the IEEE History Center and I'm here, remotely on Zoom, to conduct an oral history interview with Arogyaswami Paulraj. Can I call you Paul?
Paulraj:
Certainly
Geselowitz:
Thank you. It is November 15th, 2022. And we are conducting this remotely. I am in Piscataway, New Jersey at the IEEE Operations Center. And Paul, you are… at?
Paulraj:
I'm at Stanford University, and I live on the university campus here in the SF Bay Area.
Geselowitz:
Oh, excellent. Thank you. So, Paul, what I’d like to do is go back to the beginning and as you about your family, your childhood, and how you got interested in science and technology.
Paulraj:
Thank you Mike it’s a pleasure to do this oral history with IEEE. It’s an organization that I’ve been involved with, hmm, since the mid-1980s - a long time. To go back to my background, I was born in India in 1944 and grew up in a navy family. My father, a naval officer, moved around quite a bit on job postings. By the age of 12, I was sent to a boarding school to give me stability. That was Montfort School, Yercaud, in Southern India. I was a very strong student, fond of math and physics, while mostly ignoring everything else! Nevertheless, always and easily the topper in the class. Also, explored math on my own since the class curriculum was limited. I had heard of calculus, but there were no books on it at the school. And also learned that calculus could be used to prove that planets orbit the sun in elliptical orbits. So a lot of digging around on how calculus works followed, from limited sources – if I recall, mainly a page of basic calculus results in a book of logarithm tables. And proving that planetary orbits are indeed elliptical, of course, eluded me. I had little understanding of the opportunities in science and technology. My parents, partly motivated by financial constraints, suggested that I join the armed forces with its all-free education. Not knowing any better, I applied for and entered the National Defense Academy (NDA) in 1960. I was just sixteen years old. At the Academy, I was, again, easily the top student, and my instructors suggested that I was in the wrong career. There was no backing out, nor was I aware of better alternatives, So, the navy became my career. It was only, many years later, after joining the Indian Institute of Technology (IIT) Delhi, that I was exposed to the world of science and technology
Geselowitz:
[Interposing] I'm sorry to interrupt. Was there an emphasis though, some emphasis on science and math at NDA I know that at the American military academies, they do emphasize that.
Paulraj:
Little at the NDA. There was a lot of drilling, horse riding, and gymnastics. Some academics of course including math and physics, but it was very elementary. I think that’s changed a bit now. I opted to join the navy’s electrical branch. They functioned as technicians who maintained the weapons, communications, and sensor systems. Three years later, I graduate at the top of my class of 200-odd cadets and was sent to the navy’s engineering schools for professional training spending around five years there. The focus there was on practical skills and not on foundations like math and sciences. I recall that the credit weightage for the workshop (metalworking and like), was more than sciences (mathematics, physics). The curriculum was well below a normal BE program. During these years, I purchased and studied ME-level texts on math, signal processing, and the like.
Geselowitz:
And so how did you get to it from that, you’re in the navy, to a master’s degree?
Paulraj:
Thanks to my performance in the training courses, the navy planned for me to do a ME course, then offered at the IITs. I went for a selection interview. Fortuitously a member of the interview board was Professor Indiresan from IIT (Delhi). He was impressed and had me assigned to his institute. A few months later, in August 1969, I joined the IIT in Delhi.
Geselowitz:
So did the IIT have the reputation then that they do now as the top technical university? Was that considered?
Paulraj:
Yes. IITs already had a global reputation for producing excellent undergrad engineering students. And they remain great institutions for undergrad programs. They are not still ranked high in graduate, particularly PhD, programs. Unfortunately, at that time, the entire batch of a BE class would go to the US for graduate studies, and mostly settled there. It was euphemistically known as ‘Brain Drain’. Soon after joining the IIT, Prof. Indiresan found that I was already at the ME level and petitioned the Navy to allow me to switch to a PhD program. The Navy, understandably, was not interested and quickly declined. But, Indiresan kept persisting, and the Navy finally relented on the PhD on the condition of no more than two years of study already allocated for the ME program. By early 1970, I transferred to the PhD program, with 18 months to finish.
Geselowitz:
And who were your advisors on the faculty.
Paulraj:
It was Professor Indiresan, of course. He was a visionary professor and an advocate for indigenous technology Happily, decades later in 1997, on the 50th anniversary of Indian independence, I helped successfully nominate Indiresan for IEEE Honorary Membership. Indiresan was practically oriented and suggested the design of a pulse compression receiver for radar applications, as my PhD topic. I however preferred a theoretical subject and was rather reluctant. To my great luck, Professor Tom Kailath of Stanford University, visited the IIT in early 1970 to deliver a couple of lectures on signal estimation theory. Tom is an IEEE Medal of Honor winner and is now a colleague at Stanford. Tom’s lectures were inspiring and I began exploring estimation theory for my PhD. Indiresan generously went along. In the remaining fifteen months, I built a body of results for a credible thesis. My work offered a unified solution to the then emerging area of non-linear estimation theory, as well as expanding its scope to a wider class of problems. The work used stochastic (Ito) calculus and diffusion processes. It was very abstract.
Geselowitz:
[Interposing] So let me just ask before you resume your navy career. Two quick questions: One is, if Tom Kailath was not your advisor, did he at least serve on your committee?
Paulraj:
Tom was not my advisor but served on my reading committee. He was surprised that my level of PhD work could come from India.
IEEE
Geselowitz:
My other question is you already mentioned IEEE. So, when did you first become aware of IEEE?
Paulraj:
IEEE was not prominent at IIT Delhi in terms of student engagement and was certainly unknown in the navy. However, for researchers, IEEE journals and conferences were the gold standard. Publishing research in IEEE was the highest stamp of quality, then and so also today.
Geselowitz:
Okay. And then so now to return to your story, you returned to the navy, what, it must be now around 1971 or so.
Paulraj:
Yes, July 1971.
Geselowitz:
But you had to still write up your thesis, while you were working.
Paulraj:
Yes. The Navy was planning to send me for a tour of sea duty, but that could have killed my PhD. I still needed access to a research library, and more importantly, my very mathematical thesis needed a special typewriter, not available in the navy. Luckily, I managed to persuade my superiors to keep me in Delhi and was assigned to the Navy Head Quarters. Fortunately, the IIT, also let me stay in their guest house, which had been my home for two years. I settled into a nice routine, a day job in the navy, and nights at the IIT library, writing my thesis. That all changed suddenly in December 1971, when hostilities broke out between India and Pakistan and created an opportunity that changed my career.
Geselowitz:
Okay. So, tell me about what happened then.
Paulraj:
The Indo-Pak war began on the 3rd of December. India and Pakistan have had military conflicts earlier, but their Navies never engaged. But this time it was different. One of India’s antisubmarine ships, INS Kukhri (UK-built, Blackwood class), was on submarine hunting patrol and was sunk in a torpedo action by Pakistan Navy submarine PNS Hangor (French-built Daphne class). This happened late evening on December 7th. We lost more than 200 lives, a few of them known to me. Early next morning, the navy came knocking for me at the IIT and flew me to Bombay, the base of the western fleet. The Navy briefed me on INS Kukhri. Apparently Kukhri’s sonar (Model 170B, a UK design) had failed to detect Hangor. Sonar 170B was also widely used in the Indian fleet. I was tasked to see if its performance could be improved. I also learnt that the Navy had indeed been working with a government lab to develop retrofit equipment to improve 170B’s performance. Moreover, this equipment was in use when Kukhri was sunk. I studied the retrofit’s design and concluded that it did not add any performance value and if anything, did the opposite. It likely led to the loss of Kukhri. However, I felt confident that a lot can be done to improve the sonar’s performance. Particularly with better signal processing techniques now possible with integrated circuits (ICs). I also suggested (conveniently, no doubt) that I could do the work at IIT (Delhi) where I had good connections. The Navy agreed readily and I was assigned back to the IIT, this time as the project lead, and Prof. Indiresan was enrolled as the academic supervisor. I collected an R&D team and the IIT helped import the required digital and analog ICs. The team built a new transmitter and a receiver incorporating interesting new techniques but retained 170B’s original transducer and amplifiers. Multiple sea trials followed, and confirmed the dramatically improved range and classification performance. The Navy approved the design for fleet service in early 1973, and the design was transferred to a production agency for volume manufacture. I believe that the 170B work was India’s most advanced signal processing technology achievement at that time.
Geselowitz:
Now, was your team all uniformed officers?
Paulraj:
No. I was the only one in uniform. For the rest of the team included PhD students from IIT Delhi and a few scientists on loan from the Indian Defense R&D Organization (DRDO).
Geselowitz:
Right. Because you mentioned that the Navy was not high on letting their officers get higher degrees and do mathematics. So, I was wondering if this success changed things.
Paulraj:
The success of the 170B project changed the Navy’s perception of the potential of its own technical talent. To this day, the Indian Navy is the most technology-friendly amongst the Indian armed forces. The 170B project was also the first real payback in terms of technology innovation by the vast IIT system for the country. India had invested heavily in these institutes, only to witness entire batches of students leaving for opportunities abroad.
Geselowitz:
[Interposing] Now was the British aware that the Indians were also working on improving their legacy sonar?
Paulraj:
I don’t think so. The project was classified as secret. About 20 years later, the 170B work was described in some detail in Indian Navy’s official history. So, perhaps then.
Geselowitz:
Please continue that story. What happened? How did you get to England for a year ?
Paulraj:
Professor J. W. R. Griffiths of Loughborough University, UK, had visited IIT (Delhi) several times during 1972-73. He led a strong signal processing group at his university. He offered me a one year Research Fellowship to work on signal processing algorithms. I was just then winding up the 170B project. The Navy was initially skeptical, but seeing that I was keen, agreed. I joined Griffith’s team in January 1974. I soon met Professor Hermann Bondi, Chief Scientist for British Ministry of Defence (MOD). Bondi was interested in my PhD work, and at his suggestion, I delivered talks on my thesis at several MOD labs. These interactions gave me some exposure to UK’s signal processing R&D capability. I also met Earl Mountbatten, a name you’ll recognize. Mountbatten was the former British Viceroy to India and later UK’s First Sea Lord. Mountbatten occasionally visited the Indian High Commission in London. I was at Loughborough University then and had begun to receive a few awards from India for the 170B work. The Indian High Commissioner arranged for Mountbatten to present me one of those medals. Apart from the research work at Loughborough, I also visited sonar firms in the UK and France who had contracts to supply equipment to the Indian Navy.
APSOH
Geselowitz:
Okay. So, now, you are returning to India, and it must be about 1975.
Paulraj:
Right, 1975. I came back and immediately lobbied the Navy to support the indigenous development of a world class next gen panoramic fleet sonar. The Navy had planned to import such a sonar and was examining different options. They were skeptical about my ideas despite the success on the 170B sonar. The new project was, no doubt, far bigger. Also, a failure to deliver would have been disastrous for India’s ambitious ship building program. Happily, they finally decided to take a bet on it. I was assigned in 1976 to the DRDO’s Naval Physical and Oceanographic Lab in Cochin, Southern India, to lead this sonar program. Now code named APSOH.
Geselowitz:
So, you never did go to sea?
Paulraj:
No. The Navy waived all sea duty requirements. Building fleet sonars was more important!
Geselowitz:
Now, were you married by this time?
Paulraj:
Yes, I was. I got married in ’73, just before I went to England.
Geselowitz:
And your wife went with you to England in 1974?
Paulraj:
Yes, she accompanied me.
Geselowitz:
Okay. And that next project took seven years.
Paulraj:
From ’76 onwards, almost seven years.
Geselowitz:
Oh. And while you were on the APSOH program did you do other projects?
Paulraj:
I was full time on APSOH over all those years. I was the Chief Scientist / Engineer and Project Director rolled into one. My job was system design and verification, and program coordination with the other agencies. We worked with a number of partners, some of the major ones were: Bharat Electronics, the production agency for the sonar electronics; Mazagaon Shipyard, the shipboard installation agency, and the Indian Navy for planning and conducting the sea trials. APSOH was a dual transducer sonar, a hull-mounted array, and a second array inside a towed fish (added later). These were cylindrical arrays about 5 ft in diameter and height, with 320 transducer elements. We decided to buy the fish and towing gear from Canada, at least for the first system. Of course, all the semiconductors were imported from the United States.
Geselowitz:
By the time you’re doing building the sonar, what was the role of computers? Because they had come in.
Paulraj:
Yes, we needed a lot of computing. The signal processing, though, was too much for the microprocessors to handle. That was done by hard wired digital logic. We used a large number of 16x16 multiplier / accumulators from TRW. The sonar also needed about a dozen microprocessors for control, interface and diagnostic functions. That, however, became a bizarre tory. In 1977, when APSOH started, the Intel 8080 microprocessor looked like a good fit. The Indian government, who strictly controlled all semiconductor imports, decided that the sonars do not need microprocessors, and denied our import application. So, we had to find a way around. At that time, a computer building block known as bit-slice processors were available. Which after adding lot of additional components and design, could be used to build a full microprocessor system. The government authorities who did not understand computing, did not block import of bit-slice chips. It was a lot of extra work, but we built our computers. Pretty ridiculous!
Geselowitz:
[Laughing].
Paulraj:
But we made a bad call. The bit-slice chip we selected used PMOS technology. And the chips were very thermally unstable and would malfunction if they got too warm.
Geselowitz:
I'm guessing that they could have saved themselves a lot of time and money if they’d just let you do it the right way.
Paulraj:
Agree. Fortunately, some years later the Indian Govt. shut down the entire electronics import control department. They had done a lot of damage.
Geselowitz:
But I guess it was a good challenge for your team.
Paulraj:
It was, but it also brought a side bonus - we controlled the processor architecture and could put in lots of features like maskable interrupts and timers which the Intel 8080 did not have. But the temperature sensitivity was a huge headache. I would joke then that if you blew on the chip, it would stop working. We added additional cooling to keep it reliable.
Geselowitz:
So how did your family feel about the fact that you were in the navy, which was the family tradition, but you were working as an engineer and not going to sea? Were they happy? I mean you mentioned getting a medal. Were they satisfied with your career?
Paulraj:
I had already received many awards and enjoyed a bit of a reputation. So, they had no regrets. But everyone felt that I was in the wrong career [Laughing]. My mother, now aware of the press stories about the IIT BE /US PhD careers, knew that had I missed out. Moreover, the navy was certainly not an R&D organization and did not have a career path for me. But, for now, they showed their appreciation by giving me out-of-turn promotions.
Geselowitz:
Yes.
Paulraj:
Such promotions were unprecedented and caused me great embarrassment. I argued against it, but to no avail.
Geselowitz:
Okay. So, what happened when the project ended?
Paulraj:
APSOH was accepted by the Navy for fleet induction following a full year of sea trials on a frontline frigate INS Himgiri. It was a considerable success. Its design remains secret, 40 years later, so I cannot get into details. APSOH was more advanced than anything we saw in the Soviet Union or Europe. We did not have data on US sonars to compare. Despite the sonar’s success, the last couple of years of APSOH program saw a lot of tensions. The sonar was developed in a DRDO lab, manned by a civilian cadre, but the project lead was a Navy officer. There was a history of tensions between DRDO and the Navy, and the APSOH dynamics, which I will not go into, exacerbated this. The then the Prime Minister of India, Mrs. Gandhi, was very well aware of these damaging conflicts. Once APSOH was accepted by the Navy, I was offered a two-year sabbatical, anywhere in the world, fully paid for by the Indian Government. That looked attractive, and my first instinct was to spend that at Loughborough University. But, better counsel prevailed, and with some trepidation, I wrote to Professor Tom Kailath at Stanford. I knew him from a decade earlier in connection with my PhD thesis. I enquired if I could spend a sabbatical in his group, with all expenses paid by India. At first, Tom declined, saying that I may not be a good fit, given my decade long sonar development work, while his research group was mathematical. I wrote back to Tom a couple of times, trying to be more persuasive, and he finally relented. A few months later, in Oct. 1983, I joined Stanford, accompanied by my family.
Stanford
Geselowitz:
Okay. And how did it go when you got there?
Paulraj:
Well, it was a good break from running a large project and the incessant tensions. The Navy had zero expectations for me, beyond taking a break. At Stanford, I learnt about the work in Tom’s research group on directions-of-arrival (DOA) estimation. An area with a very long history. The underlying math discipline was linear algebra, new for me, but was easy to pick up. DOA estimation was indeed one of the key functions in APSOH. However, we could assume (after some pre-processing) that there was only one signal present. I had done experiments on sonar DOA methods for APSOH. The problem being studied at Stanford assumed that multiple signals were present, so more complex. I tried to extend my APSOH experience to this problem. Thinking intuitively, rather than mathematically, and after many months of mulling over problem, I finally stumbled on a very new approach dubbed as ESPRIT (Estimation of Signal Parameters via Rotational Invariance Technique). It was both a radical and an elegant solution to the multiple signals problem, though with mild restrictions on the array configuration. I shared the discovery with one of Tom’s students, Dick Roy. Soon, some papers, with Dick, Tom and me as co-authors, followed. A few months later, my term at Stanford ended, and I returned to India early in 1986. Dick wrote several more papers, and ESPRIT became a very well-known method, but I lost all touch. ESPRIT has been widely applied, particularly in radars. Stanford had been a relaxing and intellectually stimulating break, and my family enjoyed it too. Mixing practical intuition with mathematics turned out to be productive.
Geselowitz:
So, I'm curious, have you and Tom since then ever discussed the fact that he thought you were just a bench engineer and couldn’t join his mathematical group and then you came up with ESPRIT? [Laughing].
Paulraj:
Not really. He realized that practical experience and physical intuition can be useful for mathematical breakthroughs.
Geselowitz:
Okay. So now you were going to tell me what happened when you went back to India.
Paulraj:
Sure. When I returned, the Navy decided that I should spend more time with the DRDO, but no longer on sonar technology. The initial plan was for me to head up an existing DRDO lab for Electronic Warfare. But I finally ended up starting a new lab - Center for Artificial Intelligence and Robotics. AI was new to me. It appeared to me that useful AI was still decades out and the new lab should not get into blue-sky AI research. It was more important to build technology for the Indian Armed Forces that could be deployed in years, rather than in decades. Accordingly, I planned to develop Command Post Systems, where sensor data and weapons control were brought together for command-and-control function. And these systems could be helped by decision support, for example which weapon should fire and when. Expert systems and decision support technology was a theme in early AI those days. However, within a few months at the new lab, I was tapped to co-lead a combat jet development project, then known as the LCA (Light Combat Aircraft) program. The effort had been launched a few years earlier but was making slow progress. The government felt that since I had delivered a large system like APSOH, my skills can be used for this more ambitious program. My charter was to drive operational induction of the LCA by the Indian Air Force within seven years. This was going to be a herculean effort, even in advanced countries with an established combat aircraft industry, but I was willing to try. I spent a year, before entering the program, to understand the project’s larger ecosystem – the Air Force, the fighter pilot’s work load (by flying in the trainer versions of ex-Soviet MIG 29s), the Indian aircraft industry who were manufacturing military jets (ex Soviet Union and UK) for decades, and meeting with LCA program subcontractors in Europe and the US – for example, GE, who supplied the jet engine. I also became increasingly convinced that building and inducting the LCA into squadron service even within the decade, with a sizable Indian design content (a priority for me), and to do it functioning inside a very rigid operating culture of the DRDO was going to be impossible. I therefore proposed a more flexible quasi-government framework. The DRDO gave me the go ahead and I began to prepare the paperwork for Indian Govt.’s (Cabinet Level) approval. But alas, after six months, the DRDO changed its mind. I did not see a path to succeed, and therefore decided to revert to the Navy from the DRDO. I should add that the LCA was finally inducted into the Air Force thirty years later. It is not a frontline platform and all major subsystems from the engine to multi-mode radar are imported. India continues to import its frontline combat jets – Rafael, Sukhoi, and others. Back in the Navy, purely as an interim measure, I was deputed as Chief Scientist to Bharat Electronics, a large Govt. owned company which manufactured defense electronics like radars and communications gear. The company relied heavily on technology transfer from advanced countries, and R&D was neither their mandate nor a realistic fit for their culture in the company. In parallel with my job at Bharat Electronics, I also co-founded a Govt. program on high-speed computing. The Navy, knowing my career history, suggested that I move to the United States for a better professional opportunity. In early 1991, I wrote to Tom Kailath at Stanford, checking if he could arrange a temporary research position. But, of course, this time, he would have to fund my salary. He graciously offered me a Research Associate position for nine months. So, in October 1991, I returned to Stanford.
Geselowitz:
Now, you mentioned PM Gandhi several times. Did the assassination of Indira Gandhi affect your work at all ?
Paulraj:
Not at all, I was not at that level. Her son, Rajiv Gandhi, was the prime minister when I returned to India in 1986, but I also never met him. However, I did work closely with Rajiv’s top tech advisor, Mr. Sam Pitroda.
Geselowitz:
What happened when you returned to Stanford?
Paulraj:
My hope was that I would find a faculty position somewhere in the United States. My development experience in India was not relevant, but ESPRIT was well known and still a popular research area. Certainly, a Stanford faculty position would have been great, but a high bar. Since Tom only had limited funds, I started looking for fresh salary support. Fortunately, Gene Golub, a prominent math professor at Stanford, had a contract to study an US Air Force related signals intercept problem. And, happily he had project funds. The Air Force system used a plane mounted antenna array to separate (and listen to) mutually interfering intercept signals that arrived from different directions. Given my practical mindset, I decided to build a small experimental system with cordless phones as signal sources to try out different signal separation algorithms. In outdoor settings, things worked as expected. We could separate the signals unless their DOAs got too close. One day, however, during a rainy spell, we moved the experiment set up indoors, and noticed, to our surprise, that the signals were separable even when the phones were very close, or when touching each other. Conventional theory could not explain it. I quickly realized that signals bouncing off the walls (multipath) enabled the phenomena. This insight was new in wireless. The multipath related effect was irrelevant to the Air Force problem. But a few days later, while mulling over the results during a haircut, I realized that this phenomenon can be used to increase data rate and channel capacity in wireless links. MIMO (Multiple-Input/Multiple-Output) was born and a few weeks later a patent was filed with Tom as a co-inventor. For reasons, complex to discuss now, in the patent, the use of MIMO in wireless broadcast services was empathized, but claimed broader applicability to cellular and other networks.
Geselowitz:
So was MIMO integrated into the cellphone systems?
Paulraj:
Yes, but that took 15 years. It is now a part of all 4G, 5G and WiFi devices. Billions of people and machines benefit from wireless enabled by MIMO technology. There are many videos about those early experiments and the MIMO invention.
Geselowitz:
The video of the two students walking in the lab with cordless phones?
Paulraj:
There are several of them. Many organizations that give out awards have videography budgets!
Geselowitz:
So, while you were at Stanford did you get more involved with IEEE? I think Stanford is very actively involved maybe compared to Indian universities.
Paulraj:
I was appointed as a Research Professor at Stanford in 1993 and built a large research group around MIMO. We published in IEEE journals and conferences. I also did tutorials and keynotes on MIMO at IEEE events. And served briefly as a Distinguished Lecturer. Beginning 1993, my group ran, an IEEE sponsored, Workshop on Smart Antennas at Stanford (the MIMO label did not become popular till much later - 2005). These workshops became global hub for MIMO research and experimental results. These workshops unfortunately ended in 2000, when I began to focus on startup companies.
Geselowitz:
So that’s a great segue to the next question: How did you get into Industry.
Paulraj:
Right. I knew that MIMO works in practice, thanks to experiments at Stanford. The question in 1998 was how to break into the mainstream cellular market, which was then heavily dominated by narrowband voice services. MIMO was right for high data rate applications. There was little interest in broadband wireless for mobile internet access in 1998. The market opportunity was far out.
Geselowitz:
Sure.
Paulraj:
Therefore, I did not target mobile cellular and rather focused on Fixed Wireless Access (FWA) for home internet services. FWA technology also does not need standardization (a huge effort) essential for adoption into mobile networks. Architecturally, FWA technology was identical to cellular but did not support mobile handoffs. I founded Iospan Wireless Inc. in 1998 to build a FWA system exploiting MIMO. We married MIMO, the capacity / speed multiplier, with OFDM, a modulation technique ideal for wideband channels. Iospan built technology for both ends – infrastructure and customer boxes, and it was a lot of work. We raised more than $70 Million. But it was difficult to reach commercial sustainability and the 2001 Internet crash also dried up VC funding. Intel bought the company in 2003 for very little money. Intel adopted Iospan’s FWA technology and pumped billions of $ into the technology and the eco-system, and dubbed it WiMAX. And the IEEE 802.16 committee developed the standards for WiMAX’s fixed and later mobile services. Intel’s strategy was to build a new broadband wireless access networks, independent of the well-established wireless phone networks. This strategy began to fall apart, when the voice carriers, who had vast tower and other infrastructure in place, also began to offer (though initially low speed) data services using 3G technology. And soon, they too standardized on MIMO-OFDM technology for their next generation broadband services, calling it 4G LTE.
Geselowitz:
Okay. And then?
Paulraj:
When Intel began to develop the WiMAX eco-system after acquiring Iospan, I saw an opportunity to build semiconductors for this new global market. I therefore co-founded a new company (Beceem Communications) in 2004 to build semiconductors for WiMAX. We raised over 100 M$ in VC funding. Beceem quickly became the global leader in WiMAX semiconductors, reaching profitability. However, WiMAX ecosystem began to falter caused by the rise of LTE. The 2008 financial system crash also did not help. LTE was very similar to WIMAX, so retargeting was an option. But we now needed to support legacy mobile standards like 2G and 3G, along with 4G. This needed a lot more investment. So Broadcom Corp. acquired Beceem in 2010, using it to start their LTE chip program. This time our investors were well rewarded. I should add that by now, I had spent a dozen years juggling between Stanford and the industry and was looking for a simpler routine.
Geselowitz:
So that juggling must have created some tension.
Paulraj:
On the company side, my time spent at Stanford was seen as beneficial because of the visibility it offered me. I was the CTO, so not deadline driven. But my research group at the University suffered. It shrank a lot, and I was not raising enough research funds.
Geselowitz:
So, the students received less attention
Paulraj:
Unfortunately, yes. To make up, I began to meet them in the evenings. At home, or in my office. But it was not the same as the earlier days. Two of my students and I did a textbook on MIMO theory. That was a lot of work, but it helped lay a solid foundation for graduate students and industry R&D. My publications suffered, though my patent count soared – for obvious reasons. The industry phase did bring one nice research bonus – I always believed that one never fully understands a technology until you build it and make it work. So, building MIMO based products gave me a lot of insight and understanding into this revolutionary technology. It also gave me a more balanced picture of what is no doubt, a very complex and multi-dimensional problem. I had hoped to capture all this into another book, but never got down to it.
Geselowitz:
So, and you’re still at Stanford to this day because I see you sitting there!
Paulraj:
I am not active academically. But I do live on the campus, thanks to Stanford’s generosity. I do other things now. I am active with the Indian government on technology policy and serve on committees with the U.S. government. Help a bit with venture capital and small wireless companies. But at 79, I am clearly slowing down. I certainly would like to see the US industry build good 5G infrastructure, we are still not there.
Geselowitz:
Now is that a technological issue or just a political/economic issue?
Paulraj:
RAN Infrastructure technology for public cellular networks is complex. It has to operate across the world, in different frequency bands, spectrum allocation patterns, and support interoperability with legacy 2, 3 and 4G, and more. The traditional infrastructure vendors Ericsson, Nokia, now also Samsung, have advantages of experience and scale. A US strategy for 5G was to break up the RAN into smaller pieces under the O-RAN umbrella, to enable smaller companies to enter the 5G business. Our hope was that O-RAN model would create more competition and spur innovation. This approach, however, has created several new challenges. The O-RAN ecosystem has been slow to mature, but I am sure we will get there. On the other hand, US industry has done very well on the phone side, with Qualcomm and Apple as global leaders.
Honors
Geselowitz:
And so in 2011 you won the IEEE Alexander Graham Bell Medal for, I guess, primarily this telecommunications work.
Paulraj:
Yes, I would say for the invention and advancing MIMO research. I must have at least 200 papers and 60 patents around MIMO at that time. More importantly, I built practical systems, translating theory to practice. Making things work is hard! Mr. Bell excelled in practical technology.
Geselowitz:
A couple of years after that you became a Marconi Fellow. When did you first become aware of the Marconi Society? When they called you up?
Paulraj:
I got to know the society when it held an award ceremony nearby and somebody invited me. I did, of course, have two Marconi Fellows as colleagues at Stanford - John Cioffi and Marty Hellman. A few years later, one of my former students, and now a Professor in Europe, nominated me and I was elected fellow. Vint Cerf called with the news - that was a wonderful surprise. I currently serve on the society board. The IEEE Bell Medal is also a great honor, but there’s no family around it. But Marconi fellows are a close-knit family. Mr. Marconi was also a technology and business pioneer.
Geselowitz:
Interesting. So if I'm reading what you just told me in the past hour or so, correctly, despite all the companies and you’ve been involved with and all the big labs and everything, you really except –not even I guess the one year in England—you were technically still in the Indian Navy. So your whole career, you were either in the Indian Navy or at Stanford. Is that correct?
Paulraj:
Yes, I only have had two employers from 1961 to 2013, the Indian Navy and Stanford, perhaps a break of a 3 days to cover the India-US flight!
Closing reflections
Geselowitz:
That’s fascinating. What looks like a crazy career is actually very stable. Now, did your family follow you when you moved between India and the U.S. ?
Paulraj:
Yes, they did. I did have job continuity, and transfers and disruption are common in the armed forces. The harder part was moving to the US permanently after early retirement from the Navy. When I came to Stanford in 1991, I was already 47 years old, holding a short term research associate job. I left behind a 25-year career in India, with a record of accomplishments and many national recognitions. It was hard on my family. But in hindsight, it was a very good decision.
Geselowitz:
Did either of your children pursue a technical career?
Paulraj:
My elder daughter tried but switched to political economy. My other daughter did medicine. So, no. I am trying to convince my grandchildren to pay attention to STEM. I'm not sure that I am succeeding.
Geselowitz:
Are they in India or the U.S. now?
Paulraj:
I have two in Las Vegas and two more in London.
Geselowitz:
Okay. Great. Well, that’s really a very fascinating story. Is there anything you wanted to add that I forgot to ask you that we didn’t cover?
Paulraj:
I want to acknowledge many people and organizations who helped me over the years. The Indian Navy backed my unusual career path, an effort I know was neither routine nor easy. I also thank Stanford University for taking a chance on me, someone with no academic pedigree. Joining Stanford was the best decision in my life. It gave me opportunities at a level that I could never have had in India. Next, I am grateful to my academic mentors, Professor Indiresan at IIT (Delhi) and Professor Kailath at Stanford University. They both saw some potential and vigorously advocated for me. I owe them very much. I also acknowledge many luminaries in the US telecom industry who supported me in different ways. The most productive years in my life, in terms of engineering, was during the sonar period in India. And, in research, ESPRIT and MIMO stand out. Building practical technology around MIMO in the US faced many challenges - raising funds, hiring talent, persuading customers, and, of course, always competition. Very different from the challenges of systems development in India with its cultural and operating challenges. In terms of societal impact, MIMO easily ranks at the very top. I enjoyed my engineering endeavors and grateful for all the opportunities. The IEEE has been a big part of my life. Thank you.
Geselowitz:
Well, that’s great. So, thank you so much. I'm going to turn off the recording now. Unless are you finished? I don’t want to cut you off.
Paulraj:
I'm finished. Thank you.
Geselowitz:
Okay.
Paulraj:
Thank you, Mike.
Geselowitz:
Thank you very much
Further Reading
- Stanford University Website
- Research Pubs (400+)
- MIMO-related US Patents (total 83) See a list of MIMO patents attached
- Original MIMO Patent 5,345,599 https://patents.google.com/patent/US5345599A/en
- Marconi Prize (Marconi Society)
- Inductee - US Patent Office – National Inventors Hall of Fame
- Inductee - Wireless History Foundation – Hall of Fame
- IEEE Alexander Graham Bell Medal
- Press Release See attached.
- Video https://tinyurl.com/video-Bell-Medal
- Padma Bhushan, National Award, Govt of India:
- Citation See attached
- Weblink http://www.edubilla.com/award/padma-bhushan/arogyaswami-j-paulraj/
- Iospan Wireless Inc. Founded 1998, Founder. Acquisition report.
- Beceem Communications Inc. Founded 2004, Co-Founder. Acquisition.
