Interested in connecting with Ali Rohafza? Reach out to him on Linkedin!

About Company:

Altered Carbon is a digital vapor detection company. Their aim is to remove the complexities associated with sensors and enable users to identify the condition of almost anything in-situ, from fresh produce to humans to machinery. The same chip could detect hundreds of different things, can be updated via the cloud with more 'scent' profiles, can be manufactured via low-cost methods and run on almost zero power.

Key Questions and Topics from this Episode:

(01:34) Introduction to Ali Rohafza

(02:30) Introduction to Altered Carbon

(06:10) How does scent technology work?

(09:55) How are the sensors connecting to the scent profiles?

(11:20) What makes Altered Carbon sensors unique?

(13:08) What is the range of Altered Carbon sensors?

(15:18) Challenges to bringing vapor detection to market

(17:23) How to educate the market and customers on new products

(19:53) What are the potential use cases of vapor detection?

(22:26) Detecting drugs with vapor detection?

(23:31) Potential to create scents with data from sensors

Transcript:

- [Voice Over] You are listening to the IoT For All Media Network.

- [Ryan] Hello everyone, and welcome to another episode of the IoT For All Podcast. I'm your host, Ryan Chacon. And on today's episode, we have Ali Rohafza, CEO and Co-Founder of Altered Carbon. I have done tons of episodes over the last number of years, and this is one of the most interesting episodes I have done. Altered Carbon is a digital vapor detection company. So we're talking about smell, sensors that can detect and understand different smells. So we broke this down into everything, I learned so much. I learned a ton about how the human nose works, and how they were able to kind of break that understanding down into technology that can allow different use cases and applications in the IoT space, become even more enabled than we were before. So we're talking about what the technology is. We break down kind of where the technology's going, the challenges in building something kind of new like this, and also just the general application of how this kind of plays a role in the IoT space and in many different industries. So I definitely recommend listening to this entire episode. It's super fascinating. But before we get into that, if any of you out there are looking to enter the fast growing and profitable IoT market, but don't know where to start, check out our sponsor, Leverege. Leverege's IoT solutions development platform provides everything you need to create turnkey IoT products that you can white label and resell under your own brand. To learn more go to iotchangeseverything.com, that's iotchangeseverything.com. And without further ado, please enjoy this episode of the IoT For All Podcast. Welcome Ali to the IoT For All Show. Thanks for being here this week.

- [Ali] Thank you very much for having me.

- [Ryan] Yeah, this is definitely gonna be one of the more unique conversations with the technology that you all have developed. So I'm definitely excited to really get into it here. Why don't we start off by just having you give a quick introduction about yourself to our audience?

- [Ali] Yeah, sure. So my name's Ali Rohafza. I'm from a company called Altered Carbon. I'm the CEO at the moment, and we're working on a new technology that uses basically sensors to basically distinguish smells. So create fingerprints of different smells, and we create this library of them, and be able to trigger scenarios, and indicate what's happening. So I've been working in, with some of my business partners for years now. So it's been about five to six years and we work in the Bristol Robotics Laboratory. So we're an offshoot from the university.

- [Ryan] Fantastic. So tell me a little bit about the founding story of the company, kind of where this came from, how this even got started. I mean, when we're talking about, we talk about sensors a lot and different types of things you can track and monitor, but talking about smell and scent is something that we have definitely not covered. So I'm curious just how this all originated and kind of got to where it is today.

- [Ali] Yeah, sure. This is a funny story actually. So we were, so we were, me and my other two business partners, we were in our last year of university at UWE in Bristol, and we were basically doing an internship for this company, building a little robot, educational robots for kids. So teaching them how to code, really interesting stuff. And we constantly talked about starting our own business, and this is where all three of us met. And what we started off from was kind of trying to figure out what we really wanted to do. And one of the areas was air quality. We wanted to detect what was around us, what was happening. And so we got put in this tiny office. It was completely free so we would never complain. But we got put in this office and right next to it was a building site. So every other day we would come in, there'd be lots of like dusts like, and particles and everything in the air. And we kept complaining about it, but had no way of showing what was happening. So this is where we started to develop this technology, being like, we'll prove what's happening to you kind of scenario. And through that, developing it, we just got to understand the area much better, understand what people wanted to understand about it and not. So like a lot of interesting stuff was, the air quality is a big problem, but actually like if you just tell people that it's bad, then that's not a solution. You need to tell them also what to do and be like, okay, well, if you have these plants, or if you have this air filter, or this scenario, then it could solve your problem. But if I just said your air is polluted, then there's nothing for you to do. So that was one of our big kind of learning curves we had. So we were buying off-the-shelf sensors from other companies, putting them together in development kits and building that. And then that's where we went into building loads of communication devices.

- [Ryan] Wow, that's super interesting. So then tell me a little bit about going from there to where you are now with the role the company's playing in IoT and kind of your overall focus.

- [Ali] So yeah, where we've gone from there is, so we developed a kind of communication based products after that, 'cause we kind of didn't like how the sensors were working at the time. And we found it really hard to work with some of them. So we went, let's build the communication first. Let's figure out how to like speak from one sensor to another. So we developed a few IT products that we sold around the world. And one called Iron Link, and we developed those for a process and then kind of a really amazing, lucky accident, came up with, came up against graphene. So it's this, a wonder material, really amazing 2d material that's been kind of realized, but not actually used that much in actual technology. So loads of papers, loads of people speaking about it, but not actual commercial usage yet. So we were actually testing for batteries, testing to find out if we could make batteries out of it, and suddenly realized that actually we built sensors, and we picked up things that were going around the space. So people using certain chemicals in our labs or anything, the scent, it would pick it up, and we would see actual changes in this, the values that we were seeing. So we realized that we've made a sensor. And so we started developing that, and that's where Altered Carbon was born. And that's what we have now.

- [Ryan] Fantastic. So talk to me a little bit more about the technology itself. So when we're talking about smell, scent, different things that you can kind of build, these profiles, how is that being, I guess for a lot of people, they're probably just unsure of how that even works, let alone how this is applied and what it's used for. So maybe you kind of bring this full circle and talk a little bit more about the applications here.

- Same, it took loads of research and loads of understanding, even about biology. Like, I've had to go down so many different routes to understand it, even how it works. So, other scent, like, a lot of sensors do a similar type of work, but what we did was kind of replicate how the nose works. So you have loads of receptors in your nose, and they pick up different particles, and when one picks up one particle, and another one picks up another, it combines in your brain to make a memory. Okay? So that's how, kind of your nose, on a really basic, low down level. So I don't want any, People to come back, really low down level. That's how it works. But we have like thousands of thousands of receptors and they're multiplied millions of times. A dog has way more, like ten to a hundred times more than us, that's why it can smell more distincting, like basically distinguishing between certain things. But what we built is these receptors, each one is different and there's an array of them. And these receptors give out different signals compared to what kind of outside influence it takes. So if you imagine like the smell of someone, or the smell of a situation, or the smell of popcorn, each of them have a distinguishing factor, and each of it will hit these receptors in a different way. And we basically make a memory. So we make this digital fingerprint, which is like a memory of it and we store it. And then we have a basically algorithms that kind of track and go back to it, and go, okay, this is this, this is this. And use that as a referencing.

- [Ryan] Gotcha. And how are you determining which kind of smell profiles, or these memories that you're creating, are you just kind of building an inventory or this out of demand from use cases that you're involved in, or customers that you work with?

- [Ali] So it's a bit of both, certain parts we have interest in already. And certain parts where clients actually, coming up being like, this is our problem. So we've been, 'cause we, as I said, we were working in the gas sensing industry for probably five to seven years now.

- [Ryan] Right, right, right.

- [Ali] So we had problems ourselves and clients that would want to use some of the sensors, but had certain issues with them working in certain ways, or kind of said, I wanna detect this, but I also need this to be detected 'cause that's an issue we go through. So we kind of developed on that. But our favorite area that always came around was around detecting when food was gonna be spoiled, when it was gonna go off, 'cause it's a really big focus of ours, 'cause we think it's really important. It's like, we have enough resources to kind of feed everyone in the world. But the way it gets used is really bad. So how items are shipped from one area to another, and how it's lost. Like loads of it is lost in transit, or loads of it's lost in shelf life or anything like that. So by putting sensors in that area, we would be reducing that. And so this is where our main focus has been. But as I said, there's other people that come up and say, we also have this issue. We also need to detect if someone's transporting something they shouldn't do. Or if this chemical is exactly what it says on the tin, or can we new fingerprint something that's happening outside in the real world. So our favorite area is food, but we also are focusing in different areas and testing those scenarios, and building a database as we grow. So the more we build more of a map we'll make.

- [Ryan] So are these sensors, customizable to all these use cases, or when you buy one of these and you deploy it, it has the entire library of the different profiles that you've built in them and you can kind of, and it'll access whatever you kind of tell it to?

- [Ali] So that's a good way. So how ours work is this one chip that you may have different versions of, but the main thing is there's one chip and you can download, when you connect online, different profiles to it. And you can have like distinguishing profiles up to, at the moment, we may have like five to ten different profiles on one, but yeah, 'cause it's machine learning, the more profiles you put on a certain scenario, the more it may find it hard to distinguish between those. So if you put a thousand different profiles on it, then it may find it hard, but you can maybe have the same chip, go into an environment, say you wanna distinguish between these five or ten, and then, have the same chip downloading another profile on it, and being like, this is used for chemical sensing, and this is used for food sensing, and this is used for this. So the same chip can do all of them, but you just download different profiles on top of it.

- [Ryan] Now when you're installing-

- [Ali] And as we learn,

- [Ryan] Sorry, go ahead, go ahead.

- [Ali] No, I was gonna say, as we learn, we can add these and update clients with these new profiles, and make them just go online and say, oh, that's really interesting. We can use this for our business and they can download that profile on top.

- [Ryan] So what does the device look like, as far as like you're putting these chips into some kind of piece of some kind of sensor, right? So what make those unique compared to, I know we talk a lot about asset tracking sensors and things like that. There must be some kind of components that make it, allow it to really pick up the different pieces coming into it to match, and kind of work the correct way that the profiles are looking for the information.

- [Ali] Yeah, so imagine we use graphene. So our sensors are basically, they're really tiny and really flat, so they can go in between stuff, go into cases of stuff. They can actually be shaped into certain scenarios as well that we've been working on. And 'cause of the graphene, graphene is like a scaffolding. We add stuff in between this scaffolding, and this is what picks up some of the, kind of the gases that we want to sense in a way. So one of the unique things is we're, we use basically, virtually no power. So as a sensor itself, normally they take quite a bit of power. So you need large battery systems. So for more remote applications, it's quite hard to use. So we actually fit that need. And also because it's flat, if you have a tiny device that will fit on top of a person, like a badge or something like that, you can't have a large sensor, or on a phone, you don't have a lot of real estate to work with, and a lot of space. So you need things to be either completely flat. Well, not completely. There's always gonna be a little bit, but as flat as possible, and just larger surface areas that you can have on the back of a phone, or you need it to be very tiny, like a little bit high, but that's what you're working with. So we've built something that's, kind of fits those needs in a really good way. So you can mold it around a piece or mold it around, even this water bottle in a way to detect if there's contaminants or certain scenarios like that.

- [Ryan] And last question I have before we transition to a different topic here is, what's the range of this? So how far, or how close is this need to be to something, like, obviously when you're talking about shipping containers or storage, you can, the sensor's relatively close, but are there ever use cases where they're looking to potentially do some, or I guess evaluate something that's a little bit further away, or is there kind of limits to this, because of the way air moves and things like that.

- [Ali] Yes, so there'll be a limit to the thing you're trying to detect really. So the thing, certain gases are heavier, certain gases are lighter, so outside or indoors will change that. So certain things will basically, if the gas comes out because it's heavy or lighter, it will go up or down. And so if your sensor's put in a certain position, then it may not be able to pick it up. But what we're working on is, we've made our sensor very cost effective. So you can have a large amount of them in the same space. So in a normal sensor, you would have one in the middle of the room, like a fire detector that we have one up here. Our one, we would like, I don't know, 10 or 20 in a room. And then you'd be able to locate a certain scenario of what happens in square number 56. And go, this is where the problem occurred. This is where it traveled to, and this is where it ended up. So we're going on this different understanding of like, kind of it's called massive sensing. But I've meant as a mass sensing, as in like a lot of sensors in an environment, being able to locate certain scenarios, what happens.

- [Ryan] Wow, okay. Well, this has been a great kind of overview of what you're going, what you really have going on. I, when I first looked into it, it was, anytime you talk about scent and smell, that's something that people don't, have never really explored, and really done a lot with in the IoT space. So the fact that you are kind of venturing down this, there's definitely a ton of applications that applies to, so, that breakdown, I think, hopefully, educates our audience a little bit on what's possible. And obviously this is very new, right? And you talked a little bit about the work you had to do on the research side to really understand like how the human nose works, to kind of build from that understanding, and how scent profiles works and things like that. But just from a general standpoint, anytime there's new technologies, there's always a challenge of bringing them to market. And I'd love to kinda hear a little bit from your perspective, what some of the challenges were when you were bringing this to market to where it is now. And advice on, and I guess the idea would be for those learning, to kind of provide some good advice for people out there that maybe are venturing down a path of something that's relatively new, on how to kind of overcome those common obstacles.

- [Ali] Yes, of course. So actually a lot of our obstacles were around with sensors, definitely. It's like testing a large amount of them to see how viable it is for that sensor to be like mass produced. So we could scent, we could detect like a small amount of them. And then say, yeah, we know that it will pick up this gas, but then a client would go, well, how would this work on if we printed, if we made a hundred thousand of your sensors or something like that. So we would have to go back to the drawing board and be like, okay, so we have to design an experiment completely differently to how you first do some early testing of your work, to how you actually make it for something that would be produced and used externally. And we kind of had to develop a whole new, basically a whole new section of IP around how we test our sensors. And how we develop all this scenario of like sensors that you need to test, like thousands and thousands of elements at once in a particular way, against so many different gases a scenario. So building these training models and everything was really difficult. And we spoke to so many different people around the idea of how that would work and what they would want as clients. So what is important to them? You really need to focus on that, 'cause it's not just that what you want, kind of what the client wants and what their needs are. 'Cause sometimes some sensors, what they did was, they would just give you a standardized method and go, this is the testing we've done. And then we looked over loads to see if there's a standard method and there wasn't. So you kind of have to build it yourself, especially in our area.

- [Ryan] Right. And one thing I've thought about, and this is obviously a common question that a lot of companies battle with, is the educational component. So how did you go about educating the market, and then also customers, when you talk to them, about how this not only works, but the value it provides.

- [Ali] So we had to build our own language, honestly. So we, over time, we basically built, so certain words have bad connotations, or will say what we want to say, but doesn't mean actually around, well, we'll give people a bad idea or something. So we've had to change certain things from what we used to use to scent, or describing how we describe it. So I have this way of describing it to people is like, imagine your phone, it has a camera, it has speakers, it has, these are your sensors. And what we're building is a sense of smell for your digital devices and for your computer. And that took a very long time, even though it's really simplistic, but like to get to that simple way of explaining it to people that everyone will understand, compared to I could, back in the days, explain it in a more scientific way, but doesn't mean everyone would understand what we mean. So it's like the language is one of the hardest things to focus around your market and who you're actually going for and like who you're focusing on.

- [Ryan] Well, oh, I will say that through this conversation, you've been able to very clearly articulate, kind of how it works. And I think, I'm sure, that's probably even come a long way in, through your research, and how you found what resonated with people from even just a general sense, because even just friends say, hey, what do you do? And you're trying to explain to them, you're building sensors to detect different smells and scents. I mean, I'm sure, that just opens up a whole different can of questions, so.

- [Ali] Exactly. With their understanding, will ask completely different question as well. So that's really interesting.

- [Ryan] A hundred percent.

- [Ali] They'll poke holes in areas where you have to learn like, wait a second, we didn't actually think of that. That's really interesting. But that's why it's good. We have a really diverse team, so. Each of our team, come from like sometimes different backgrounds, which is really good. Some from robotics, some of them from embedded system, to business, to communications and media. And so. Together we can come up with solutions that if you had a team that all thought the same wouldn't. So sometimes we can help each other and benefit from a somewhat, like, sometimes someone doesn't understand coding, but they could still help a coder sort of thing.

- [Ryan] Totally get it. So one of the last questions I wanted ask you before we wrap up here is just talk about, this is a new area of focus for not just the podcast, but I think for the industry as a whole in many senses, no pun intended, but what does the future of this look like? Like where does this go? How does this get involved with, maybe, in the medical field, with infectious diseases, and people, I'm sure it plays into the agriculture field, where do you kind of see this going and see this evolving into over the coming months to years? Just cause it feels like it's a new area that is, it could be very popular.

- [Ali] So that's really good, 'cause we're basically looking for partners in these really interesting areas, you mentioned as well. It's like vertical farming, agricultural areas, a really interesting one. 'Cause we imagine in the future, we can detect things from pesticides to illness in certain plants like fungal infection, stuff like that. That basically decimate crops. To even detecting like fires before they actually occur. So like you can detect if your computer starts getting hotter than it should. And it's about to like go off, like there's all these indications of smells being released before it actually goes, so instead of waiting until fire happens, you can actually indicate way before. We imagine this technology being spread out through so many different industries, to actually basically learn from your environment, and learn from their surroundings, and to be input with data, like, being like a reward based system where it learns, oh, there's an anomaly. You should maybe have a look at this. And then you say, oh, was wrong or right. And this could be implemented in so many areas. So we did a project with UWE on detecting like different infections and wounds, and these bacteria aren't bad bacteria, but when they get over a certain amount, so like they grow to a certain amount, then can become infectious and they can become bad. And so just detecting those and doing those situations, we hope in the future that we could use it in military, medical applications, food, so many different areas. You can imagine it could be put into any sort of environment that uses smells or has that indication. There's a project someone did that they can detect. So they have dogs that can detect diabetes and cancer just through smell. So if you imagine we could potentially train our scenario to anything that a dog could do in a similar environment.

- [Ryan] So, then I guess, I was actually gonna ask you about dogs as a follow up question, when you have like drug sniffing dogs and things like that, is that something that their noses are more sophisticated to be able to pick those up? And that's something that it's, is that possible for a sensor to be able to replicate and also do at maybe larger scale, if they can't have endless amount of dogs doing drug sniffing at airports and different areas?

- [Ali] Yes, definitely. It is definitely something the technology can do. I think, that the focus we have is more around the food area, and that's our main focus, but yes, of course, the sensor can do majority of the stuff. So scent-based stuff like detecting drugs, anything that releases a smell. They that's how they pick it up. They smell those particles in the air. And they can pick that up, and we can be trained to do that. A dog can only pick up a couple, they can be trained to pick up multiple, but at a time when going out, they'll give them an item and they say, find me something similar to that. So we would be doing that similar training scenario.

- [Ryan] Gotcha. This is maybe a little bit of a tangent here, but so what about the opposite side of this? And this is something I think, I'm sure people have thought about, is there a way to create something, and this may not be directly connected, but because you're building these scent profiles and you understand the makeup of them, is there a way to kind of have this go the opposite way and create something that pushes out a certain scent? Based on, like through technology, like your computer, you're reading something and like, ah. You're watching a cooking show and they can make it smell like what the show is actually, like what they're cooking in the video.

- [Ali] So there would be, in the future, there'll definitely be a way of translating it. And there's a few companies over the R&R radar that basically make these, basically scents in bottles. It's like a tincture sort of style scent, that are like mixture of certain flavors, like blueberry and lemon and stuff like that. And you could, in theory, as a base material, build something that would like spray a few different scents and mix them together to make a profile that you could have maybe on your desk, on your computer. And we could take it in from one end of the signal and then translate it to the device. Yeah. So that's definitely a future idea t