An Interview With the Inventor of USRP: Happy 10th Birthday

Matt Ettus, founder of Ettus Research, answers a few questions on the 10th anniversary of the USRP platform.

SANTA CLARA, Calif. – January 16, 2015

The USRP turned 10 years old this year, but let’s start from the beginning. How did you get into software radio?

As a kid, I was always into computers and electronics. I eventually got into HAM radio and radio communications, so radios were always an interest of mine. I went to Washington University in St. Louis for undergrad and then to Carnegie Mellon for my masters degree.  At CMU,  I concentrated partially on computer architecture and partially on wireless communications—my masters’ research focused on wireless and communications. I researched adhoc networks – metropolitan-scale ad hoc networks.  I then I went to work at Sun Microsystems designing microprocessors.  My next job was at a GPS company, and that’s when I got back into wireless comms.

I first heard about the software radio concept when I was in grad school attending a conference. I was presenting some work about what you could call a software radio, although I hadn’t even heard the term at the time. I eventually started several hobby projects related to software radio.  While I was working at a Bluetooth chip company in 2001, the GNU Radio project started, and I joined up right away.  GNU Radio is an open-source software radio library that offered all users the fundamental building blocks to build software radios.

And you’ve continued to stay active in that community?

Yes, since long before I started Ettus Research, I’ve been contributing to GNU Radio and that’s really what the company is all about. As the company has grown, we’ve continued to contribute code to the project. All the software we write is open source and we see ourselves as important members of the community.

The community is interesting because you see a wide variety of technical questions.  For example, we get questions as specific as, “how do I get this to compile?” But, we also get questions from a lot of people who are learning wireless communications and are learning wireless comm through experimentation with GNU Radio and our hardware.  Those guys ask a lot of systems questions.  By being part of the community, we get to help educate people and teach. One of the things I like is when someone is building a radio using GNU Radio using our hardware, and they have questions like, “why is this feature implemented in this way?”, or “is this an artifact of the hardware, the software, or what’s the theory behind it?” I get to teach and to share, and I really like that part of it.

At what point did hardware start to become a part of the community?

Well in the beginning, we were doing very interesting things.  We made an HDTV decoder back around 2002 or so, but right, there wasn’t any interesting hardware to play with. No matter what you put in software, you still can’t plug an antenna into your USB port to see if the code you’ve written is actually working. You need some device to connect in-between, and it really didn’t exist at any sort of reasonable price level. I had come up with the idea for USRP and I was presenting it at conferences trying to convince people to make the thing. Nobody did, so I started to build it myself. Eventually, I found some funding from the National Science Foundation through the University of Utah.  So I was able to quit my day job and work on it full-time. That was August of 2004, but even then the intention wasn’t to sell the devices. In theory, I would just design the hardware and support the design so that others could build the radios.  However it just didn’t work out that way.  So, I ended up building the hardware too and then selling the radios myself.  Eventually we grew, and I hired people to help, because up until that point I’d been doing all of it.

Really, I never intended to make a company out of it.  It was always “I just want this product” and no one else was making it, so I had to build it myself.  I was customer number one, and so I had to make it.  And it turned out, luckily, that other people wanted it as well, so it turned into a business.

Now that USRP has been around for 10 years, what are some of the best applications you’ve seen, both professional and hobbyist? 

We have a lot of people doing wireless communications research—spectrum monitoring and surveillance applications are represented well in our customer application base.  Some of our more interesting applications have definitely come from hobbyists. We’ve had a lot of people do security research with our devices. People have proven that certain wireless standards are completely insecure, like for example, wireless communications from a control to implantable pacemakers can actually be hacked and that was demonstrated with our hardware. People have used our radios for all sorts of things, from radio astronomy to wildlife tracking. For a while, people were tracking muskrats using our system. They basically implemented a  direction-finding systems to track where the muskrats were located.  You may have seen a guy on television from National Geographic walking around with headphones and an antenna in the wilderness …  The USRP and GNU Radio automate this formerly manual process, and electronically replicate that application.

People have used USRPs to build cell phone base stations to provide cell phone coverage at Burning Man, in underdeveloped countries, on islands out in the Pacific and in Africa where there’s no other infrastructure. We’ve had people use them for medical imaging, and some people have even used USRPs for sonar work which is really neat.

There was a group at a university that came up with a system using the USRP to use gestures within your house using the WiFi signals that are already present. Just by looking at how your body affects the WiFi signals, they can tell what gestures you’re making through walls and throughout the whole house.

What are you focused on developing right now?

Well our newest generation products have much, much bigger FPGAs. It’s a real focus for us and we’ve created a new architecture for what goes inside the FPGA that we’re calling RF network-on-a-chip, or RFNoC. It’s aimed at making it easy to use these massive FPGAs, because you have these incredible resources, these huge FPGAs that Xilinx is making now that scale with Moore’s law.  At the same time, your ability to program these devices doesn’t get easier but in fact becomes more difficult. So we need to make it easier for people and that’s a key focus for this most recent generation of products going forward.

What technologies are on the brink of coming to fruition right now that you’re really excited about?

There have been a lot of important new wireless techniques over the years that people have developed the theory for, and have then gone on to demonstrate the theory on our hardware, which has always been really exciting to me. To see things like interference alignment and cancellation demoed for the first time on our hardware, is great. So I think right now, some of the most interesting research is in the area of massive MIMO, where a user will have many, many antennas in a system. You might have 128 or more antennas in a system, and that has very interesting possibilities for cell phone communication and other wireless communications, but it’s also extremely hard to do at a hardware and a computational level, and so a vast majority of the work is really theory right now, and people aren’t building systems to really use it. So, I’m excited to see people like Lund University in Sweden use our hardware to demonstrate practical 100-antenna MIMO systems.


About Ettus Research

Ettus Research™, a National Instruments (NI) company since 2010, is the world’s leading supplier of software defined radio platforms, including the Universal Software Radio Peripheral (USRP™) family of products. With an overall affordable system price, expansive capabilities and fully supported by the open source community, USRP products are used by thousands of students, engineers and scientists worldwide and remain the top choice in software defined radio hardware for algorithm development, exploration, prototyping and developing next generation wireless technologies across a wide variety of applications.