Thomas Watteyne
Photo by Scott Brammer
.Thomas Watteyne is an insatiable enthusiast of low-power wireless mesh technologies. He is a researcher at Inria in Paris, in the new EVA research team, where he designs, models and builds networking solutions based on a variety of Internet-of-Things (IoT) standards. He is Senior Networking Design Engineer at Linear Technology, in the Dust Networks product group, the undisputed leader in supplying low power wireless mesh networks for demanding industrial process automation applications. Since 2013, he co-chairs the IETF 6TiSCH working group, which standardizes how to use IEEE802.15.4e TSCH in IPv6-enabled mesh networks, and recently joined the IETF Internet-of-Things Directorate.
We asked him about his work, his ties with California and the 2 projects that made it to IPSO Challenge 2015 semi-finals that will take place in San Jose, California, on December 1st..
.
.
– Thomas, can you tell us about your research?
I work on what we call the Internet of Important Things.
These are networks composed of tiny battery-powered wireless devices which form a multi-hop wireless mesh network.
The technology we develop and standardize allow those devices to live for years of batteries while offering ultra high reliability.
People typically use them for industrial applications (such as wireless control loops) but also for Smart Building and environmental applications.
I joined Inria in January 2015, and together with 2 other faculty, we started the EVA Inria Project Team; we are based in the Inria-Paris research center.
.
That sounds exciting!
It certainly is!
It’s exhilarating to work in a field with a real pull from our industrial and research partners.
We have established strong partnerships and secured funding allowing us to build the dream team to tackle this challenge.
.
– What are your ties to California?
I used to live and work in the Bay Area before joining Inria.
I did my postdoctoral research at UC Berkeley in Prof. Kris Pister’s team, the father of the “Smart Dust” vision.
In late 2010, I joined Dust Networks, the startup company which is market leader of industrial wireless sensor networking solutions.
Inria has really allowed me to keep strong ties with California, something which was very important to me.
I still work part-time at Dust Networks, which I see as the “best of both worlds”: I lead my research team and work with my students at Inria, while keeping strong ties with the Silicon Valley.
At UC Berkeley, I co-lead the OpenWSN project with Prof. Pister’s group, the de-facto open-source implementation of the networks we develop, and in January 2015, we started the REALMS associate team between Inria-EVA, Prof. Steven Glaser’s team at UC Berkeley and Prof. Branko Kerkez’ team at the University of Michigan.
So in the end, I feel like I work somewhere above the Atlantic Ocean [laughs].
.
– Tell us more about the REALMS associate team
It’s a truly exciting project.
Over the last couple of years, Prof. Glaser’s team has deployed 945 sensors all around the Sierra Nevada to monitor the snow melt process.
California has a drought problem; these sensors gather unprecedented amounts of information about the environment (snow level, soil moisture, solar radiation, etc), allowing hydrologists to build accurate models about where the water goes once the snow melts.
This helps the State of California deal with the drought problem in a very efficient way.
The great thing is that the teams in REALMS are very complementary.
The Berkeley and Michigan teams are absolute experts at exploiting the data received, and have gained incredible practical expertise at deploying those networks.
At some point, bears pulled on the wire connecting the wireless device to the sensors, so now the cables are Kevlar enforced.
This is not something you can learn in a textbook!
And we know about the networking aspect, as they use the technology we are developing, so it’s a pure win-win situation.
.
– What project are you currently working on together?
Carlos Oroza, a graduate student researcher from the Berkeley team is currently visiting our team in Paris.
Together, we are applying ensemble regression-tree techniques, a form of machine learning, to the connectivity data we have gathered on the deployed networks.
Over the last year we have collected some 40GB of data on which device can hear which other device, and at what reception power.
The tools we are developing allow us to predict how well a device will connect to the network even before we go in the field.
This means that, rather than spending precious time moving repeater nodes around to find the positions which give the best connectivity, we will let our computers crunch on the numbers for a couple of minutes and install the device at the best location at the first attempt.
.
– What do you expect from this collaboration and what would be the next step?
Research is all about collaboration: collaboration with teams in different countries and with different backgrounds.
And REALMS is exactly that: we get to teach each other tools and techniques we would never have used before, collaborate on papers and projects together, and have a great time.
In 2015, we took 4 trips between the teams; this has allowed us to lay a very solid base for our future collaboration.
We see the Inria Associate Team opportunity as the perfect vehicle for collaborating.
But we also see it as a form
of seed money; in 2016, we plan to apply for further funding to start a larger project.
Our goal is to apply this technology which now “just works” to problems in Europe: in environmental, smart building and smart factory applications.
We’re actively reaching out to local partners as we speak.
.
– To change topics slightly, I’ve seen that two projects proposed by the Inria-EVA team made it to the semi-finals of the 2015 IPSO CHALLENGE. Congratulations! Can you tell us more about the challenge and your projects?
Thanks! Yes, we’re truly excited!
IPSO is the “IP for Smart Objects Alliance”, a group of about 50 companies which develop technologies around smart objects which connect to the Internet.
The Alliance is a great way of promoting the technology and encouraging collaboration between the companies, but also with the academic and standardization worlds.
For a couple of years, IPSO has also been organizing a yearly challenge in which people submit ideas using Internet-enabled Smart Objects.
A team of experts reviews the submission, and the ten semi-finalists compete during the finals; this years, the finals will be on Dec 1st in San Jose, California. We truly believe in our projects.
.
Our first project is called “micro plug and play” (uPnP) and is a collaboration with Prof. Danny Hughes’ team from KU Leuven in Belgium.
His team designs a middleware solution which allows you to connect a physical sensor to a wireless device, and that device will detect what sensor it is, download the appropriate drivers from the Internet, and automatically send sensor data into the cloud.
This allows Internet of Things type applications, for example analyzing the movement of people in a shopping street, without requiring any programming knowledge.
Just stick the device to the walls, insert a sensor and see the sensor data “live” on your cellphone screen.
.
The second project is called HeadsUp!, and applies “Smart Object” technology to the medical world for helping patients recover from a retinal detachment eye surgery.
One effective surgery to treat retinal detachment is called pneumatic retinopexy; its downside is that it requires the patient to keep his/her head in a specific position for 1-3 weeks.
The HeadsUp! project proposes an IP-enabled Smart Object which revolutionizes this post-surgery experience: it continuously and wirelessly reports the position of the head to a computer, which emits a warning directly and/or through other IP-enabled objects when the head’s position moves outside of its the recommended tolerance.
.
– It looks like you’re having fun!
Are you kidding, I’m the happiest person alive! [laughs]
/
*** Interview by Tania Castro, European & International Partnership Department, in Paris on October 23th, 2015.
For more information about the topics discussed:
– the OpenWSN project: http://openwsn.berkeley.edu/
– the REALMS associate team: https://realms-team.github.io/
– the IPSO challenge: http://challenge.ipso-alliance.org/
– the uPnP project: http://www.micropnp.com/
– the HeadsUp! project: http://www.headsup.tech/
