Transcription
New types of sensors are about to flood the marketplace, but are they also going to disrupt how we live our lives and in the workplace? We're going to check in on what technologies are driving this new wave of sensors on this episode of Today in Tech.
Hi everybody, welcome to Today in Tech. I'm Keith Shaw. Joining me on the show today is Greg Rouse. He is the president of Asahi Kasei Microdevices, or AKM. Welcome to the show, Greg.
Wow, that was a mouthful. Sorry about that. Talk to me a little bit, Greg, about who AKM is and your role within the company.
Yeah, so AKM is a division of Asahi Kasei. Asahi is a large Japanese conglomerate, and AKM is the Microdevices division of that. As you mentioned, one of our primary focuses is on sensors.
Alright, so when we talked before the show, there were a lot of different innovations happening in the sensor space, and we haven't really talked about that on the show lately. I wanted to ask you about, within the last few years, what is causing this disruption? Is it just a case of everything being smaller, cheaper, and faster, or is there something else going on that you're seeing in the marketplace?
Yeah, I mean, I think one of the driving factors is that the automotive industry today is requiring a lot more sensors. So that drives the volume, it drives a lot of innovation, and now that innovation that was done for automotive is migrating, for example, into people's homes.
That's really allowed for a price point to come down where it can really affect a lot of people's lives because of the innovation of cross-platform or cross-industries from automotive to consumer or home-type products. So we're seeing a blending of automotive, medical, and consumer now with these kinds of sensors.
Was it primarily driven by the move towards autonomous vehicles, or was it also just the driver assist portions of the automotive space?
Yeah, I mean, I think initially the industry was really pushing for autonomy. I mean, ten years ago, they told me we were going to be driving autonomously in five years. Now, we're ten years later.
But I think that kind of the end goal is what drove the innovation, where the cars would be autonomous. That was kind of the end goal, so that drove a lot of development and a lot of investment into the industry.
But that turned into driver assistance or driver monitoring because there are a lot of steps between a car from two or three years ago to get to fully autonomous. So those steps along the way required several technologies to be developed.
There are milestones that you can go along the way, and then those milestones that were developed for that technology can be applied to varying industries. So that's kind of where you're seeing customers or companies pivot to take the technology developed for the autonomous vehicle and apply it to a variety of different industries.
Right, so that's where I think you've already answered my next question. Do you see these new types of sensors hitting the market, or is it more about existing or older technologies that were applied to something like the automotive industry and are now just being developed in new ways?
Yeah, that's correct. I mean, a lot of the innovation to get it into the car was to be able to get the price down, shrink the geometry of the semiconductor, and get the volume up. That enabled the lower cost.
What's interesting, though, is we're seeing some of the tech that was slated for the car actually coming out in consumer home products faster because automotive takes longer to develop. It's a slower design process.
Okay, so even though the tech was originally done for automotive, we're seeing some industries actually adopt it faster than automotive. But the price point that was originally targeted for automotive is what drove it.
You know, it's all interesting. Do you see new types of sensors that make you think, "Oh, I didn't know you could do that with a sensor?" Or has the world of sensors reached a point where there's nothing that can't be sensed these days?
I mean, I think historically the problem has always been, you know, like the most interesting thing of the recent tech, which I think is really driving the market, is actually edge AI.
The advantage of edge AI is that you now have enough processing locally that you could put a sensor in a room, a sensor in the car. This edge capability gives you the ability to process it locally and send a small amount of data.
That way, you can because AI algorithms require a lot of data. A lot of that's basically what differentiates AI from a standard algorithm: you give it tons of data, and it's able to capture anomalies even below the noise floor.
So the ability to process that locally and have some sensor stream data constantly to it, I think that's really some of the technology that was available sensor-wise in the past but wasn't being utilized because there wasn't edge AI capability.
That's something I think that, even in automotive, even if it's a giant computer they're putting in a dash, technically that's edge AI because it's not happening in the cloud.
Yeah, so that's really what's enabling these sensors. Having edge AI capability is enabling a whole bunch of different applications for the sensors.
Yeah, I remember in the days of the Internet of Things and the IoT sensors that were coming out, the big selling point was always about having the sensor in a remote office that's 50 miles away.
You know, it would be a small amount of data that goes across, so you don't need a heavy network anymore. But as you know, those applications were limited because you can't have a lot of data crossing the network. If you did, it would be too expensive, and you would need faster networks and things like that.
But now you're saying that with edge deployment and possibly 5G, you can look at new applications for a lot of these sensors that you couldn't do before. Is that fair?
That's correct, yeah. That's correct. And then there are a bunch of AI training engines and stuff that are coming out, enabling algorithms that weren't really possible or people didn't think could be developed on the edge before.
So that was definitely one of the trends I've seen recently: a lot of AI training companies are training algorithms that enable AI to work better for you or quicker training or what data you give it.
That's kind of one of the things that really helps enable these markets.
Alright, so let's go over some of the new markets that you're seeing. You gave me a list ahead of time that I want to dig into a little bit.
You're saying that aging tech is one of those big areas where you've got a lot of people getting older and wanting to stay in their homes. How has that technology helped by new types of sensors?
Yeah, I mean, to give it some context, all Boomers will be over 65 by 2030. So if you think about it that way, that's only five years away.
That means there's going to be a significant population of people that still want to live in their homes. They don't want to be disrupted, and they're also not necessarily scared of tech because a lot of Boomers have been using smartphones for the last ten years.
So they're a little more accustomed to it than someone who's already in their 80s. It may be more challenging for them to understand some of the technology.
I think that's part of the reason the industry has put a lot of effort into aging tech: one, because of the aging population, but also because they're a little bit more tech-savvy than the previous generation.
A good example of that is millimeter wave. For example, millimeter wave was designed primarily for autonomy in the car. You could use it for a child left behind in the car or for autonomous driving.
So inside the car, it was 60 gigahertz, and that 60 gigahertz can also be used inside people's homes. You can do things like fall detection or motion tracking.
You can tell if someone is laying in bed versus laying on the floor. For example, you can detect someone that's fallen without invading any privacy because historically, the alternate solution was to use a camera.
People don't want a camera in their bathroom or bedroom. The advantage of using millimeter wave is we can get the outline of someone falling.
The way we're differentiating with millimeter wave is you can basically detect the breathing object. Millimeter wave is so accurate we can tell that that object's breathing.
That's how we can tell a human versus a chair, for example. If you fall down, we can tell how long you've been there.
Some concepts that came up are like, for example, say your mother falls down. You would get a notification that, hey, your mother had an event; she fell down.
You call her and say, "Hey, how are you feeling? I see you may have fallen down." She says, "Oh, I'm fine. It wasn't a big deal."
But then the following day, she stays in bed all day. Clearly, you know that she's not okay, that she's in pain, and she should be checked.
There are a lot of statistics that show if you break your hip, the faster you have surgery, the shorter your recovery. Those are the kinds of things.
A lot of people are hesitant to go see a doctor, so you have someone else on the outside that could encourage you to go in based on changes in events or lifestyle.
You can also use millimeter wave for tracking prescriptions, for example. You could train the system to know they take their pills at this desk at this time.
You could set notifications if they didn't follow their pattern that day. There are all kinds of things you could do with millimeter wave without violating any of their privacy.
Yeah, well, what I love about millimeter wave is that initially, the technology was trying to be used to promote 60 gigahertz for Wi-Fi and home wireless.
It was all about needing this speed for VR stuff, and of course, the VR technology didn't really take off. I love that now millimeter wave can be used within the home but in a different scenario or use case.
That's a great example of how a technology can follow in new use cases instead of the ones that it was initially promoted for.
Correct, yeah, exactly. A lot of it's just really interesting because there are several companies doing the same kind of thing. They've developed it for automotive, and they're doing consumer versions, going after aging tech.
It can also be used in medical applications, like in hospitals, finding if someone fell out of their bed, for example, or if someone got up out of their bed that wasn't supposed to be moving.
You can notify a nurse to go check on them if they're not supposed to be getting out of bed. So there are a lot of applications that it could be used for.
Alright, and so I want to move on to a couple of other use cases. I think we were talking before the show about smartwatches and smart wearables.
You've seen some temperature sensors, some sensors that can detect fertility cycles, glucose monitors, things like that. Talk a little bit about how that's changed over the last few years.
Yeah, I mean, I think the main difference with temperature monitoring is that there have been several wearables that have monitored temperature.
But this year, we've launched a new sensor that tracks it basically measures with a gap, so it doesn't have to have contact. It's a non-contact temperature sensor.
The advantage of that is if you're wearing a wearable like a watch or headphones, if there's movement, it doesn't need contact. It can basically detect the surface of the skin.
So you can monitor temperature while you're exercising, while you're walking, just during normal activity. Several other devices today only monitor your temperature accurately while you're sleeping or motionless because they're using a contact type.
Part of the reason they're doing that is they're trying to maintain battery life and things like that. This newer sensor that we came out with just gives you more flexibility to track your body temperature over time.
One of the use cases for that is actually to monitor the cycle of a woman. That way, you can know when you're more apt to get pregnant, for example.
You can track the fertility cycle of a woman trying to get pregnant. That's an interesting use case where you can get more accurate data if you're monitoring it all the time.
Do you find that a lot of the applications around the sensors depend on the software and the application of it? Or do you have to switch out the sensor that might already exist?
For example, if you're developing something like a glucose monitor or a smartwatch that can track temperature, do you have to switch out the sensor, or can you use existing sensors but just use them in new ways based on the software?
Yeah, I mean, I think like for example, the sensor we launched, its differentiating factor is it doesn't have to be still or in contact with the skin.
Because of that, it enables new algorithms. Now there are different use cases and different ways you can utilize that sensor.
As the sensor technology improves and becomes less cumbersome or more power-efficient, that will allow a lot of new applications to open up.
Alright, and I've got one more interesting one that I want to talk about, but then I've got a fourth thing that I want to bring up on the energy side of things.
Chemical and liquid detection: you've seen something like a smart diaper being developed using a chemical or liquid detection sensor.
This thing is, again, so all my kids are beyond that phase of their life, but a smart diaper that can alert you when it detects chemicals in there. I think we all know what chemicals they're detecting.
So it's, again, back to an old principle. You know, like when you were in elementary school or junior high, you took a lemon and poked in two little rods, and then you measured the voltage across that.
I don't know if you remember doing that, but essentially that creates a very small voltage when you do that reaction with a lemon or potato or however you did it back then.
If you put a polymer and an electrode in the diaper, now we can do the same principle. When that diaper gets moisture in it, we can use that with one of our energy harvesting devices, which can boost that power to a usable level and then transmit a wireless signal to your phone or home IoT network.
So you're basically making a smart diaper that is battery-free. The advantage of that is you don't have any processing going on, so it's very safe.
There's no electronics that are running all the time. The electronics don't turn on until the diaper needs to be changed or someone needs to take some action against the diaper.
Yeah, and so there are two use cases. One use case is potentially you could change your infant even before they woke up. You could try to change their diaper before they started crying.
The other option would be for the elderly. Some elderly people are staying in bed, and maybe certain nights they have issues, certain nights they don't.
You want to go check on them without waking them. This would give you an option where you could utilize this technology to know if they had an accident or something like that.
It could also help with nurses. The other advantage, too, is this electronics can just be transferred from one diaper to another.
So the electronics themselves would essentially last forever because it would just click onto the electrodes.
Is something like this more impressive not just because of the sensor stuff? Because I think you and I joked about, well, I've got a pretty good sensor for a diaper, and it's right there between my glasses.
Or is it more about the energy savings? Because again, that's the other thing I wanted to bring up: you can now take a lot of technology that will allow devices to harvest their own energy.
Can you explain a little bit about what you mean by that?
Yeah, so there are several different approaches. You can harvest energy from wireless now. There's so much Wi-Fi in people's homes now.
If you've got a low-power sensor, you can basically harvest that wireless energy, store up a local charge, and then take a measurement or transmit something if you need to.
There are also solar panel technologies that have progressed dramatically. Now there are indoor solar panels.
From a relatively small solar panel, you can do things like window sensors or a TV remote control that would never have a battery. It would just capture interior light and charge it.
The advantage of that is it goes with the green initiative that everyone's trying to do. People are trying to be carbon neutral, and throwing away or recycling batteries is definitely one of the challenges in the world today.
Using these energy harvesting technologies, you can make devices that don't have batteries. It's better from a user experience too because you don't have to worry about changing batteries and things like that.
I think from the diaper perspective, that's one of the best advantages: you don't have to worry about, "Oh, did I change the diaper? Is the battery in the diaper?"
It's kind of like you never have to worry about it. It's just always going to work based on the reaction.
Those are the kinds of things that I think will enable these sensor technologies to be more part of your life because they're seamless. You don't have to worry about maintaining them.
It'll just happen automatically, and I think that's what will really enable these new applications as people want to age in place or want these sensors to become part of their lives.
Like you brought up the indoor solar sensor too. I'm old enough to remember those solar-powered calculators that we all had. You would shine it at a light to charge it up, and then you would use it.
You wouldn't have to put a battery in it. I'm sure those things still work if I can go dig up a calculator somewhere.
Do you think those will make a comeback too as this energy harvesting definitely happens?
Definitely, because the amount of energy you can get out of those now is much higher. That will enable a lot more.
The other thing is as sensor technologies' power consumption goes down and as these other alternative energies or energy harvesting methods increase their efficiency to capture them, you're going to see more and more.
Like that old calculator you had back in junior high.
Yeah, just watch out for big battery. That industry is big in the EV market, so don't worry about that. They have plenty to sell to now.
Now, you brought up a couple of cases when we were talking about the aging in place issue where there is always a potential for false positives.
You brought it up from the privacy perspective with cameras, but I wanted to get your opinion on this. There was a high-profile example, I want to say last year, maybe even before that, where people wearing Apple Watches were going downhill skiing, triggering the accelerometer or whatever sensor was on there, thinking that they had gotten into a car crash.
There was a big article in the Wall Street Journal about all this. How does the industry try to address that false positive factor?
You don't want to start making sensors that are going to go off all the time and give you false positives. I want to really make sure that my mother-in-law has fallen rather than just dropped the sensor that she was wearing or whatever was detecting that.
Yeah, and I think one of the advantages is, like in your home with AI, one of the ways you can do it is by tracking their patterns. For example, you know that she normally doesn't wake up at a certain time and lay down and stretch at a certain time.
You can also track. From that particular example, one of the advantages of having the device off of your body is that you can see how they fell.
Even if it's just their overall shape, you can get a velocity of how they fell. Did they lay on the ground for a certain period of time?
When you're wearing a wearable like a watch, all you can really tell is an impact. Whether you slam it up against the wall or the ground, in that example on the screen, they say, "Oh, are you okay? Did you fall?"
But when you're skiing or something, you've got it covered up, and you don't see that message. You get a false positive that way.
So I think edge AI applications can really help to alleviate the false positives, but eliminating false positives is basically mandatory for the market.
Yeah, because anytime you get false positives, the users will just ignore it, and it's useless.
Does that also mean you have to be linking the sensor to other data tracking? You want to make sure that the data is tracked, and you've got the software that understands the data.
It's almost like these devices will become data trackers in addition to just the sensor part, right?
Yeah, so a lot of companies are making kind of an ecosystem. There's not really a standard yet, but there's going to be an ecosystem of these devices where they'll interlink.
It'll track your movement outside when you're wearing your wearable, and then it can track your movement inside your house. But it'll all be seamless.
I think that's really the big advantage. Then you'll just be notified for anomalies: your heart rate, breathing rate, or why your temperature is higher today.
Are you feeling okay? Those kinds of things. You get a lot more preventative alerts.
Telehealth was a big trend during COVID, but I think a lot of this sensor technology will be able to alert you for things that are going wrong in your life or things that could change the course of your life earlier on.
It'll try to help prevent some of the chronic diseases that so many people have as they get older.
Who are the leaders in the space when you talk about the application layer of a lot of these new devices? Is it dependent on the industry you're looking at?
Like aging in place is going to be different, or aging tech is going to be different from maybe health and fitness?
Yeah, I think that's the case because a lot of these big companies have their own niches and customer bases.
But a lot of them, one of the things we're doing is trying to branch out. For example, you may want to wear a wearable for health and fitness, but then let's say you're a female trying to get pregnant.
That's kind of the same market, even though that's more health combining with wellness. Those are the kinds of things that I think you'll see more of: those combining applications will be the next step.
Do you see, so that makes me think of another question. In the future, are people going to have multiple devices attached to different parts of their body, or is it going to be a single device that will track a lot of different things and have a lot of different sensors on it?
Or is it going to be a mixture of both, depending?
I think, in my personal opinion, it's going to be a mixture because it depends on your activity.
If you're at the gym, you may just want to wear headphones. If you're in your daily life, you may want to wear a smartwatch.
Or if you're in the house, you may want to wear nothing. You want to take a shower.
It depends on what will make this tech really popular: when you're not thinking about it because the device is just a part of your life.
If it's too clumsy and cumbersome, it's not going to be popular. I think that's really what's going to enable this.
I think we're at an inflection point because now you can make devices that are a lot more seamless.
That's really why these devices didn't exist ten years ago.
Do you see a lot of the big tech companies exploring this, or is this more of a startup space where you're just seeing individual new companies?
We see it in startups, big tech, and medium-sized companies. Everyone's targeting it, and the main thing is because they can take their existing business and expand it.
It's a way to add value, and a lot of people want to get into the data space.
You have the people that own the clouds interested because there's more data to be pumped into the clouds. You have the people in the middle who want to make devices that can feed information and give feedback and add value.
Then you have startups trying to come up with new applications for it. It really fits all of them, actually.
Is it going to be advantageous for the big tech companies that will eventually win out, or does it just depend on the market?
It depends on the application, but you may see big tech companies do acquisitions as well. Startups with good ideas are definitely in a large market, so that's why so many companies are focused on it.
Alright, and you were at CES this year. Was there anything there that jumped out at you that you looked at and went, "Oh wow, that's a neat idea?" Or did you even get a chance to walk away from where you were probably meeting with customers and stuff?
Yeah, I mean, our booth was incredibly busy because of the smart diaper. That generated interest from even people that were in unrelated industries.
I think it really opened people's minds to other applications where you can do a smart device that has a sensor-based device that has no battery and feeds an edge AI device.
From my personal view of what I saw there, edge AI was a lot more realistic now.
There are a lot more companies making processors at varying price points and processing levels, depending on what you're trying to do.
You don't have to buy a square peg and try to fit it into all these other applications. There are so many edge AI processors available, and that's going to really enable a bunch of other applications.
I think that was the most interesting aspect to me because we were pushing sensors, but a lot of people were linking them with other things they were making.
From my perspective, that was the most interesting thing at CES. There's definitely a lot more aging tech talk this year than in the past.
I think automotive was maybe reduced some at CES—not as much buzz. Someone came up to me and told me, "Hey, you can only show a bigger monitor in automotive so many times."
A bigger screen here, more screens there. There was less change in the automotive market, so there wasn't as much buzz around it.
But there are always trends and waves in the tech fields at CES.
Well, it definitely feels like everyone had AI on their mind. I was going to say, is AI a big component in this space as well?
I think you already answered that question when you talked about edge AI.
Yeah, I mean, it's what it's enabling. Without AI, it's kind of not possible to get the performance you need.
Alright, so now, for those not familiar with the market, how long would it take from some of these new announcements where you have these types of products and use cases?
How long does it usually take companies to develop this where an average consumer will then be able to see the stuff on the market?
Is it a matter of multiple years? Is it just within a year? Is it really faster than that? How fast can companies move from concept to delivery?
I think what you'll see is that companies will launch concepts within a year, and then they'll go to full production.
Most companies are trying to target a two-year design cycle or less, just from an investment perspective, trying to capture the market trend.
In automotive, it's more of a three to four-year development time. That's why I think some of these technologies that were originally for automotive you'll see in consumer first, just because the design can move faster.
Yeah, some great stuff, Greg. Again, thanks for giving us an update on what's going on in the world of sensors. I appreciate you being on the show.
No problem, great talking to you, Keith.
Alright, that's all the time we have for today's episode. Be sure to like the video, subscribe to the channel, and add any thoughts you have below.
Join us every week for new episodes of Today in Tech. I'm Keith Shaw. Thanks for watching.