How Japan’s Unique Relationship with Robots is About to Make it #1 Again

Japan had been a global leader in robotics for decades, but recently the traditional Japanese leaders have been losing ground to the better-funded and better-publicized firms coming out of America and China.

Mujin is changing that. While iRobot and Boston Dynamics have been grabbing headlines and YouTube views, Mujin has been quietly breaking ground with a series of real-world commercial successes in deploying the next generation of industrial robots.

Perhaps Mujin's largest achievement to date has been their project for Chinese e-commerce giant JD, in which they developed the world's first fully-automated logistics warehouse where robots unload the trucks, stock the shelves, and them pick and pack the items for shipment without human intervention.

Today we talk with Issei Takino, who founded Mujin with his co-founder Rosen Diankov, and he explains why Japan looks at robots in a fundamentally different way than Western countries do, and how that will lead to a significant competitive advantage.

It's an interesting conversation, and I think you'll enjoy it.

Show Notes

How to get the ecosystem to adopt your platform Why robots have not yet taken over industry (or the world) How to get your first customers in robotics How to get feedback from reluctant Japanese customers When being a Japanese startup is an advantage How America and Japan view robotics and automation differently Advice for starting companies with multi-cultural teams The critical differences between Japanese and American universities

Links from the Founder

Everything you ever wanted to know about Mujin Friend Issei on Facebook See Mujin's robots in action Video of Mujin's automated logistics warehouse (this is very cool)

[shareaholic app="share_buttons" id="7994466"] Leave a comment Transcript Welcome to Disrupting Japan, Straight Talk from Japan's most successful entrepreneurs. I'm Tim Romero and thanks for joining me.

Today we're going to talk about robots, specifically industrial robots.

Now Japan has been a global leader in robotics for decades, but in recent years, Japanese firms seem to be losing ground to the better funded and better publicized companies coming out of the US and China.

Well today we're going to sit down with the founder of a company that is already starting to change that. Issei Takino founded Mujin with his co-founder Rosen Diankov and they have developed a kind of android for industrial robots, that is to say, it's a generic operating system that works with almost any hardware and works far more effectively than anything else in the industry.

Issei and I go into some of the details during the interview but perhaps the clearest illustration of Mujin success was a project, they did for Chinese e-commerce giant JD. They developed the world's first fully automated logistics warehouse. It's a massive facility but almost no humans work there. Robots unload the trucks, stock the shelves, pick the items for delivery and then pack them and ship them out. It's hard to explain in an audio podcast, so check out the video. We've got a link at the site and it's really amazing to watch.

Issei and I also talked about how Japan and the West look at robots very differently and how that might be holding America back.

He also shares his experience and advice about founding and running a start-up as a multinational team and we talked about why these kinds of Japanese foreign partnerships are going to become more common and more important in the coming years.

But you know, Issei tells that story much better than I can, so let's get right to the interview.

[pro_ad_display_adzone id="1404" info_text="Sponsored by" font_color="grey" ]

Tim: All right! So I'm sitting here with Issei Takino of Mujin, the maker of controllers for industrial robots. So thanks for sitting down with me.

Issei: Oh! It's my pleasure, thank you, thank you for inviting.

Tim: Well I'm glad after all the technical difficulties that were finally able to sit down and talk. Why don't you tell everyone, what exactly the Mujin controller is and why it's important?

Issei: So basically like an illusion control eyes, it's like Android and like I know IOS in cell phone industries. Right now like, industrial robot, it's actually needed with the highest demand ever. However you know necessarily what is very hard to use

Tim: Right.

Issei: And then like UI, you know, the interface is everything different. Every maker you know everything different but every maker, so it's like a harder to use and the robot is not smart, that's why the application is limited.

Tim: So the Mujin controller is basically an operating system for industrial robots that, sits on top of the basic hardware.

Issei: Yes.

Tim: So to run on any manufacturers robot.

Issei: Yeah, yeah exactly.

Tim: That's fantastic, I can see why that is great for people, who want to develop applications for industrial robots or the buyers of industrial robots.

Issei: Yeah.

Tim: But aren't the manufacturers worried about this, I mean don't they look at you and say, wait a minute, this is Mujin is doing to us, what Microsoft did to PC’s, they're making it commodity hardware.

Issei: Yes, of course, the equipment makers worry. I had to get confidential information from each maker and that was hard in a very beginning of the negotiations. In 2013, there was only one maker you know, which is Denso Wave. Gradually more decided to open up their interface.

Tim: What was the driver for that because there are so many startups, who have an idea similar to Mujin and debt me the idea that, they could provide a generic interface?

Issei: Yes, yes.

Tim: So what was the trigger, that made the manufacturers, stop resisting and start using the Mujin controller?

Issei: So the basically the uniting the interface it's not enough. That's not enough motivation for them to the open ¥ their confidential information, right? They try to close their systems, applications, the control language everything. And we'll be trying to open it. So this is already a conflict of interest, right?

Tim: Right.

Issei: So in order to make them open, there are conventional information. We need a killer application, right?

Tim: So, so what was your killer app?

Issei: Our killer app is and it's called a Bin Picking.

Tim: Bin Picking?

Issei: That's Bin Picking, so called.

Tim: Right and bin picking is been a, one of the most challenging parts of factory automation for a while because it's, well why don't you explain, why it's so difficult?

Issei: So bin picking is very scary. You know, you know there's a two-part, so why is that can a 3D vision has to detect the parts and then like the robot has to move it, have to take a motion to pick out. So one is detection and the other is motion. So for human beings you know pick up from in a randomly positioned part is super easy.

Tim: Right.

Issei: It's like a natural like a, for the robot, for the robot it's super difficult for them to do it.

Tim: Because you could have a bean of parts that are in, all were you too in different directions.

Issei: So let me explain with the robot, I have to control in this robot. The current way is you have to teach. So usually, you teach the robot and the robot just repeats. Maybe you can sit, as long as you have a very good 3D vision, maybe they can detect the parts but after you detecting the parts, the robot, it has to be taught, so you have to remember, what has to be taught, right? The orientation and the positions everything! It's almost impossible to know.

Tim: Every possible permutation would be overwhelming.

Issei: Yes overwhelming and it's impossible to teaching we have before because there are so many possibility of all the motions, right? Motion planning is like a basic technology, which makes a robot think, how to move you know the automatically.

Tim: So it's motion planning more teaching robots to be goal oriented rather than perform specific motions is that exactly?

Issei: Yes, yes exactly, with all the technologies even though you can detect the parts but you still have to teach, you have to teach there are all the possible motions by yourself. So it's almost impossible to teach. Well with our system as long as the robot cab see, it can move. You don't have to teach.

Tim: Right. Okay, well listen before we dig into the technology and competition and the market in general, let's back up a bit and talk about you?

Issei: Okay.

Tim: You started losing back in 2011 with your co-founder Dr. Rosen Diankov, how did you two meet?

Issei: I'm sorry everybody misunderstood that like, actually I convinced him to now join the Mujin but this is so wrong.

Tim: Okay.

Issei: So actually, I like, we met in Japan. At that time he was working at Intel and Willow Garage, which is one of the most famous robot startups. So he was like super like the opposite character of me. Super different, right? So he graduates the best in his class at UC Berkeley and after that, he entered a Carnegie Mellon University, which is the most famous for robotics.

Tim: Yeah.

Issei: And then, he got PhD, when he was 26 years old and after that, he was working at Intel and Willow Garage, and Microsoft and those company. At a time they came to Japan for the exhibition and they had a booth and well you know the very smart like genius people. There's only one Japanese guy in Willow Garage, and he was my mentor in business.

Tim: All right!

Issei: Yeah so he asked me to help out because I, you know, I was working in an Israeli company and I was top sales and he knew I can speak English and also like Chinese and I know the industries, right? So he asked me to come to help. He was like superstraightforwardd guy like, once he believes, he doesn't,