Stellar Lab Radio #5 Guests：Mianxiong Dong and Kaoru Ota
In this episode of Stellar Lab Radio, we are joined by Mianxiong Dong and Kaoru Ota from Muroran Institute of Technology. Their research aims to realize communication infrastructure that remains resilient even during disasters. Guided by their unique network philosophy, “Ten–Chi–Jin” (Heaven–Earth–Human), they work across disciplines including information and communication technology, wireless systems, and medical-engineering collaboration.
In the first part, we explore how their research journey began with their personal experiences during the Great East Japan Earthquake, which led them to reflect deeply on the essence of human connection. We discuss their vision for disaster-resilient communication networks using drones and satellites, as well as their cutting-edge research on brainwave (EEG)-based interfaces and even the reproduction of dreams.
We also delve into their research philosophy—how they balance personal curiosity with real-world implementation—and their “group-based” research structure, where multiple labs collaborate and evolve together.
In the second part, we will trace how they each chose the path of becoming researchers, and how they arrived at their current interdisciplinary approach.
From the personal experiences and values that shape their work to their vision for the future, we hope you enjoy the upcoming second part as well.
Listen from: here
The Origins of Tenchijin and the Experience of 3/11 — How He Learned His Mother in Minamisanriku Was Safe
Kita: Today’s guests are Mr. Dong and Ms. Ota. Thank you so much for joining us.
Dong: Hello.
Ota: Thank you for having us.
Kita: So, it’s been snowing quite a bit today, and you’ve come all the way from Muroran — thank you so much for making the trip. How does it compare to Tokyo?
Dong: Well, Tokyo is definitely warmer.
Kita: I heard from Ms. Ota that the plane was grounded at the airport?
Ota: Yes, we landed at New Chitose Airport right on the day it was all over the news, so we were surprised by the enormous crowds.
Kita: Right, they were saying it was going to snow yesterday, even here in Tokyo. I was worried since I ride a bike, but it turned out not to snow much. (laughs)
So, let’s dive right in. For the first half, I’d love to hear about your research. You both specialize in information networks, IoT, and disaster communication. To help our listeners get a clearer picture, could you walk us through what you’re currently working on? Let’s start with you, Professor Dong.
Dong: Sure. I have two main topics, and one of them is disaster preparedness. The reason I got into this field goes back to 3/11 — the 2011 earthquake. At the time, I was studying abroad in Canada. When the earthquake struck, it was the middle of the night in Canada, and when I woke up and turned on the TV, I saw that something terrible was happening in Japan.
Kita: Right, there’s a time difference.
Dong: Exactly. It was daytime in Japan, but the middle of the night for us. I found out about the earthquake through the news, and my first instinct was to contact my parents — they lived in Miyagi Prefecture, and the tsunami had caused devastating damage.
Kita: Of course.
Dong: I tried to reach them immediately, but I couldn’t get through at all. Meanwhile, the TV kept showing news about nuclear issues, towns being completely wiped out by the tsunami— my parents lived in Minamisanriku, which was almost entirely submerged and destroyed. I was extremely worried, but I just couldn’t reach them. It took three, maybe four days before I finally made contact. That became one of the key turning points that led me into this research.
After returning to Japan and graduating, I first joined NICT — the National Institute of Information and Communications Technology — and then moved to a national university a year later. But that experience I had in Canada never left me: the desperate need to reach the people you care about in a disaster, to confirm they’re safe. That stayed with me, and when I joined the university, it became the driving force behind choosing this field of research.
In other words, I wanted to take the network and communications research I had been doing during my doctoral studies and apply it to disaster situations. There are many approaches to disaster research — civil engineering, information science, medicine — but I wanted to contribute from within my own field, and to draw on that deeply personal experience. That’s where it all started.
Kita: I was in Japan at the time, so I was able to reach people fairly quickly. How long did it take you?
Dong: It was really about three or four days before I could confirm my family was safe.
Kita: Nothing at all during that time?
Dong: Nothing. To be precise, I was able to speak with them by phone after three or four days, but in the meantime I tried every possible method to find out if they were okay. The phone infrastructure was completely down — especially for international calls from Canada. That applied both to landlines and mobile phones. The internet infrastructure, however, recovered relatively quickly, so I turned to that. Twitter was just emerging at the time—
Kita: Twitter?
Dong: It was still very new. I actually signed up for the first time — I still use that same account today — and I started searching for people and trying to get information about my family out there—trying to get information through what we’d now call retweets.
In the middle of all this, there happened to be a photographer. My parents worked at a resort hotel, and this photographer was there too. He took photos of the disaster area and uploaded them to the internet — but not from the disaster zone itself, since all infrastructure there was down: no electricity, no internet. Instead, he walked about 90 kilometers to Sendai, where services had been restored, and uploaded the photos from there.
Using Twitter and other tools, I was able to spot my mother in one of those photos— and that’s how I confirmed she was safe. I wasn’t able to actually speak with her until much later, when the phone infrastructure was finally restored.
Kita: You found her in a photo?
Dong: Yes, exactly.
Kita: Wow.
Dong: It was a miracle, really.
Kita: It truly was.
Dong: Yes. Minamisanriku was one of the hardest-hit areas — I was terrified my parents had been swept away. I had no idea what was actually happening on the ground. But photos have timestamps and location data embedded in them. The fact that a photo taken after the tsunami showed my parents’ home — with them in it — meant I could confirm they were alive.
Kita: Incredible.
Dong: That was one of the key starting points. And when I think about translating that into research, it comes back to the story of the photographer. He couldn’t upload the data on the spot, so he physically carried it — walking all the way to Sendai. That became the inspiration for a system we developed called “Tenchijin,” a disaster-resilient communication system.
Kita: That’s the one developed in partnership with companies, right?
Dong: Yes. The core idea was: how do we build a network that allows fast, affordable communication when disaster strikes? “Tenchijin” stands for the network of the sky (ten), the ground (chi), and people (jin). The “people” network is exactly what that photographer embodied — a human physically carrying data from one place to another. That real-world action inspired the concept, and specifically gave us the idea for the “jin” — the human — layer of Tenchijin. That’s how the name came about.
Kita: When you mentioned he walked 90 kilometers, I initially thought maybe the photos were geotagged and that’s how the location showed up — but that wasn’t the case at all.
So after the earthquake — after that experience with disaster communication — you mentioned you’ve continued researching in that area. Is that still your main focus, or have other fields started pulling your interest as well?
Connecting from the Sky: Drones and Satellites Opening Up Disaster Communication Networks
Dong: The communications field is really the backbone of Tenchijin — it has several components, but communications is at the core. Professor Ota has been advancing that side of things through various JST-funded projects. Professor Ota, would you like to share more?
Ota: Sure. Professor Dong talked about how his personal story led him to the Tenchijin research, but my path also shifted during my time studying abroad in Canada.
I originally specialized in information networks within computer science — specifically how data travels across the internet, which is structured like a mesh, with many possible routes between a sender and receiver. My research focused on how to optimize those routes. But my host professor in Canada was an authority in wireless communications, so my research started shifting in that direction. Wireless communications is somewhat different — it’s less about routing and more about how to make efficient use of radio waves between a device and an access point. So I gradually moved my research axis toward wireless communications, which is where I am today.
To come back to Tenchijin — when the concept was first developed,satellite communication wasn’t widely available yet, so one of the key ideas was to deploy drones as flying access points to build a network. Now, of course, we have Starlink and satellite communication has become available even for home use, giving us global coverage. During last year’s Noto Peninsula earthquake, for instance, ground-based networks went down and Starlink terminals were deployed in the disaster zone. But satellite communication still has its drawbacks — it requires open sky to work, and direct communication capacity is still somewhat limited. So we believe that using drones to cover areas, as proposed in Tenchijin, remains a valid and effective approach for rapidly building emergency networks in disaster situations. We’re continuing to research that.
Kita: When I think of drones, they’re machines — just like mobile phones — so aren’t there limitations in terms of battery life?
Ota: Absolutely, there are significant constraints. The drones we typically use can fly for around 20 to 30 minutes under normal conditions. But when we ran an experiment in very cold weather — around 1°C — it took a long time just to get the drone started, and even once it was running, it only stayed airborne for about 10 minutes.
Kita: I’ve always wondered why the cold drains batteries faster. Is it the same reason your phone dies quickly when you’re skiing?
Ota: Exactly the same — it comes down to lithium battery’s behavior in the cold.
Kita: So does that mean you need to look for alternatives?
Ota: It’s one of the challenges we’re working on. The startup phase consumes the most power, so one approach is to minimize unnecessary movement by optimizing the drone’s flight path — what we’d call UAV route planning. Automating that planning process is one of our current research topics.
Kita: I had always thought of Tenchijin as primarily Professor Dong’s project, but it sounds like you’re both deeply involved. Are you running this as a shared group, or did you start out separately and gradually converge?
Dong: Going back to our university days — I completed my undergraduate, master’s, and doctoral degrees at the University of Aizu, while Professor Ota did her undergraduate there , then went to the US for her master’s before returning for her doctorate. Our labs actually overlapped at certain points — we were in the same lab as undergraduates, went separate ways for our master’s, then Professor Ota came back to the same lab for her doctorate. So in a sense, we’d been working on similar research themes since our undergraduate thesis days.
When we both joined Muroran Institute of Technology, we had independent labs but collaborated where it made sense. A large-scale system like Tenchijin falls broadly under communications and information networks, but it really requires multiple research threads to function — and we each have different strengths and weaknesses, so we’ve complemented each other. Since then, my interests have shifted somewhat toward medical-engineering collaboration, while Professor Ota has continued pushing forward on the communications and networking side.
Kita: Was that shift toward medical-engineering a natural evolution from your disaster research — moving into medical support during disasters, for instance?
Dong: Partly, yes — though it might also come down to personality. For me personally, research has to be interesting first and foremost, and on top of that, it needs to be practical. Of course there are many fields of pure basic research, but in computer science — surrounded by things like smartphones, AI, and network infrastructure — the applied side tends to dominate. Theory research is a smaller part of the picture, at least in my perception. So within that context, thinking about how my research can be put to practical use has always been my starting point.
Moving Objects with Brainwaves, Replaying Dreams — The Expanding World of Medical-Engineering Research
Dong: Tenchijin grew out of a lived experience of not being able to confirm my family was safe. The medical-engineering work comes from a similar place. Super-aging societies are something Japan, China, and much of the world will increasingly face, and frankly, I’m getting older too. (laughs) In that sense, I want to bring my expertise to bear in ways that intersect with my own life . Being able to weave together personal interest and research — that’s really one of the great privileges of working in academia, of being a professor. That’s why I’ve become so drawn to medical-engineering collaboration lately.
Kita: You mentioned the appeal of university research — but I’m curious, since computer science is also actively researched in industry, what keeps you at a university? What makes Muroran Institute of Technology the right place for you?
Dong: The way I see it, university research is fundamentally idea-driven. You’re allowed to pursue things that may not generate profit anytime soon — that’s the nature of academic research. In industry, there generally needs to be a visible endpoint, a return on investment — maybe not always, but in most cases. At a university, if something catches your interest, you can just go for it. You can start a project simply because you’re curious. . That’s a big difference.
Kita: That makes sense. When people in life sciences talk about “implementation,”they often think of applied research or things like drug development.. But even implementation-minded university research is different from industry — it’s more idea-first, curiosity-first, rather than working backwards from a defined outcome. That’s the value, I think. So concretely, what does your medical-engineering work actually involve?
Dong: We are working on several areas. One is a collaboration with Sapporo Medical University, led primarily by Professor Ota, where we’re developing AI algorithms to assist with urological diagnosis. Beyond that, we currently have several doctoral students working on three themes. The first is medical digital twins — using simulation to model individual patients’ health data and predict outcomes on a personalized basis.
The second involves EEG data — brainwave signals — and using that data to control objects.
Kita: Wait — control objects? What do you mean?
Dong: Think of a computer mouse. We’re exploring whether you can move the cursor using brainwave data alone.
Kita: Are you wearing something on your head?
Dong: Yes, exactly — you put on a device that captures the EEG signals, and by analyzing that data, you can do things like move the cursor to the right.
Kita: I had no idea that was possible.
Dong: That’s one example. The third theme is something a student of mine is currently working on — recording EEG data continuously during sleep, over a full eight hours. We all dream, right?
Kita: Right.
Dong: And sometimes you wake up knowing you dreamed something, but you can’t remember what it was. What this student is trying to do is capture those brainwave signals during the dream state, and then reconstruct what was dreamed. So if, say, you dreamed about drinking coffee, that would be recorded in the EEG data — and even if you wake up having forgotten it, the goal is to extract that segment and replay it: “Last night, you were drinking coffee.”
Kita: That’s incredible.
Dong: Honestly, when I first heard it, I thought — that’s genuinely fascinating.
Kita: Dreams can be so wild and unpredictable — the storylines, the way scenes shift. You could use them for creative writing, or just to revisit a really wonderful dream you’d otherwise lose forever.
Dong: Exactly. It’s a research topic I find genuinely exciting.
Kita: It seems like there’d be a natural collaboration with sleep researchers, no?
Dong: Absolutely. We’re still at the data analysis stage, just getting started — but I do have connections in the sleep research world. If anyone listening happens to be a sleep researcher… (laughs)
Kita: Fingers crossed! And the researcher who came up with this idea — that curiosity itself is remarkable.
Dong: That’s the heart of it, really. Research starts with a wild or unconventional idea, and then you methodically test it, piece by piece — that’s the most enjoyable part. In our lab, we practice that. With so many students bringing so many different ideas, the default is: “Sure, let’s try it.” And when something interesting comes out of it, we ask: “How do we push this further? Where should this go?” That’s how we operate.
Ota: EEG data is getting a lot of attention right now, but it’s genuinely difficult to work with. There’s even research on implantable devices — electrodes placed directly in the brain — particularly in the US. But that’s frightening, isn’t it?
Kita: Very frightening. I’d worry about it just… dissolving in there. (laughs)
Ota: Exactly. So we use non-invasive headset-type devices to capture EEG data — but the tradeoff is significant noise in the signal, which makes data processing quite challenging.
Kita: I’d imagine implanted electrodes give much cleaner data?
Ota: They do, yes — much less noise.
Kita: Going back to the mouse control — there are so many everyday moments where you want to give a command but can’t move. For me, it was putting my kid to sleep and desperately wanting to signal to the next room to be quiet, without actually saying anything. If I could just think “open the door” and it would happen — that would be amazing. Is that actually possible?
Dong: With the right mechanical components attached, yes. We already have voice control — “Alexa, stop the music,” “Alexa, turn down the air conditioning” — but the problem is you have to say “Alexa,” which wakes everyone up. (laughs) The next step would be capturing EEG signals and linking them to those same devices — doors, air conditioning, lights — without saying a word. I think that’s entirely within reach.
We decided to build a team, not a lab with silos
Kita: That’s fascinating. Do you have ideas about taking this even further — building on your current research, multiplying it across new areas?
Dong: Computer science moves incredibly fast. So much of what exists today simply didn’t exist two or three years ago — smartphones, VR, AI. The defining characteristic of this field is that you have to keep updating yourself, keep engaging with what’s new. Given that, we combine our own curiosity with a constant push into new territory — medical engineering, disaster response, communications — just keep challenging ourselves with new things. That’s really the ethos of researchers in this field.
Kita: At Stellar Science Foundation, cross-disciplinary exchange has always been central to our vision — the idea that new thinking emerges from collision between fields. And computer science, I’m realizing, might be especially well-suited for that. I hadn’t really thought about it field by field before. Have there been moments in our community where you’ve encountered an idea from another discipline that genuinely shifted something for you?
Ota: Nothing comes to mind immediately, to be honest.
Dong: I do think computer science is exceptionally well-suited for cross-disciplinary collaboration. Data analysis has become so central to research across virtually every field, and wherever data analysis and algorithms are involved, computer science naturally overlaps. At the various events I’ve attended, I’ve found myself thinking: “We could probably work together on something” — across multiple conversations with researchers from quite different backgrounds.
Kita: That makes sense. On a related note — you’re both PIs running independent labs, but you also operate as a group of PIs working together. That’s not a structure I come across very often. Is it intentional, or specific to Muroran Institute of Technology?
Ota: At Muroran, at least in the information science department, assistant professors can become PIs from the moment they join. When I arrived as an assistant professor, I already had my own independent lab. The benefit is obvious — you can set your own direction and lead your team from early in your career. But larger groups allow you to combine strengths and cover for weaknesses. We never sat down and said “let’s form a team of three” — it just organically came together that way.
Kita: So rather than each lab being siloed, the looser group structure means the PIs can stay in constant communication and keep updating each other as they go?
Dong: Exactly. In a university setting, students are constantly coming in and graduating — that cycle happens every six months, really. Fresh thinking flows in, people grow and move on. The group expands and contracts naturally. What holds it together through all of that is the core — the core technology, the shared vision, the soul of what we’re doing. As long as that’s solid, the group can come together, drift apart a bit, and come back together again. Professor Ota joined in 2013, I arrived in 2014, and we started with one faculty member and four students. Twelve years later, we’re around fifty people.
Kita: Fifty people across the whole group?
Dong: Students, faculty, staff — all combined, around fifty. It’s grown and shrunk at various points, but the overall trajectory has been upward.
Kita: In industry, headcount growth is often treated as a milestone in itself. But in research, is there an ideal size,particularly in computer science,where a team just works best?
Dong: Great question. My personal opinionis that too big is just as problematic as too small. In an academic context, one hundred people would be genuinely difficult to manage because I want to be involved in every project, every discussion, every progress check. That level of engagement has a natural ceiling. So we’re not aiming for that. I think we can grow a little from where we are now,what do you think?
Ota: I’d agree. We also have sub-clusters within the group organized by research theme, and we rely on doctoral students in leadership roles to manage those smaller units.t’s not realistic for faculty to oversee everything directly.
Kita: Do you each have a preferred management style? Hands-off, or more closely involved? And do you differ from each other?
Ota: I tend to give people room to work. I’m more of a watch-from-a-distance type, honestly. That worked well when the group was smaller, but as we grew I realized I needed to step in more. Still, my natural preference is to observe and let things unfold.
Kita: Where did the difficulty start to show as the group got bigger?
Ota: More people means more variety — and more interpersonal dynamics. Relationships between people multiply. That’s just human nature, I suppose.
Kita: Professor Dong, what about you?
Dong: I also believe strongly in giving people the freedom to follow their own ideas — that’s a non-negotiable for research. But in terms of where the management approach has to change, my gut feeling is around twenty people. Let me give you an analogy. Having one child versus two versus three are fundamentally different situations. Take a family trip: with one child, you add an extra bed to the hotel room and you’re fine. With two or three, one room doesn’t work anymore — you need two rooms, and suddenly that search on the booking site gets a lot harder. Same with cars — five passengers and six passengers require completely different vehicles. It’s the same logic with research groups. Having watched our group at every size as it grew, twenty feels like the real inflection point — the moment when the nature of the thing changes. We’re at fifty now, and I think we can stretch a bit further with our current setup. But if we ever reached a hundred, we’d need a fundamentally different system.
Kita: So roughly sixty or seventy is where you’re aiming?
Dong: That’s my sense, yes. And the way I think about it — it’s not about total headcount across multiple sites, but about each physical group as its own unit. Twenty is one kind of system. Twenty to fifty or sixty is another. Each location has its own logic.
That’s all for Part 1 of our episode with Professor Dong Mianxiong and Professor Ota Kaori. I hope you enjoyed it.
In Part 2, we’ll go deeper — into what first drew each of them to a life in research, the experiences that shaped them, and the passion that continues to fuel their work today. Be sure to check it out.
See you soon,on stellar lab radio.
Stellar Lab Radio – exploring world-changing research that nobody knows about yet.
Listen from here
