Subscribe
Share
Share
Embed
Brains, Black Holes, and Beyond (B Cubed) is a collaborative project between The Daily Princetonian and Princeton Insights. The show releases 3 episodes monthly: one longer episode as part of the Insights partnership, and two shorter episodes independently created by the 'Prince.' This show is produced by Senna Aldoubosh '25 under the 147th Board of the 'Prince.' Insights producers are Crystal Lee, Addie Minerva, and Thiago Tarraf Varella. This show is a reimagined version of the show formerly produced as Princeton Insights: The Highlights under the 145th Board of the 'Prince.'
Please direct pitches and questions to podcast@dailyprincetonian.com, and any corrections to corrections@dailyprincetonian.com.
Senna Aldoubosh 0:00
Hi everyone! Welcome to Brains, Black Holes and Beyond, a collaboration podcast between the Princeton Insights Newsletter and the Daily Princetonian. My name is Senna Aldoubosh
Speaker 2 0:23
and my name is Ria Tomar. Today's guest is Atsutse Kludze, a graduate student in the Electrical and Computer Engineering Department. He graduated from Cornell University in 2021 with a degree in Electrical and Computer Engineering and he is now working in the smart wireless agile networks lab also known as the Swan lab. His research interests lie primarily in wireless communication and sensing devices. Atsutse, welcome to the show.
Speaker 1
Thank you.
Speaker 2
So, for our first question, what got you initially interested in ECE?
Speaker 1 0:51
So, what got me interested in ECE was based on the motto that my undergrad at Cornell had, which was that ECE's could do everything. I always wanted to do engineering and STEM, but what got me is that I want to always have the background, always the technical skills to basically do anything I want in the sense that if I want to explore one avenue was say of research or even like this, any private industry, I can always maneuver between them. And as you know, ECE students have a lot of skills that are very much well-valued in basically any form of life you want to go through. So let's say you want to go for something like programming, or you want to go into something like chemistry or science or physics. ECE is fairly the basis as it kind of allows you to do all that at once and even then you kind of craft what you want to do. Like taking positive skills was from the computers portion or medical portions to basically have that kind of flexibility that I think that most majors did not allow.
Senna Aldoubosh 1:51
Awesome. And so like, while we were researching for this episode, we kind of stumbled across like a an article talking about like a new technology produced in the ECE department specifically in the lab that you're in about using 6G wireless signaling to detect the quality of produce and how that could have like implications for the reduction of food waste, specifically. So could you walk us through like, well, first of all, like, what is 6G wireless sent signaling and like where did the idea to apply that to like produce come from?
Speaker 1 2:19
Yeah, so 6G wireless sensing distance is basically the new future network. So you know, for example, your phones use 4G or 5G, they still use for example six gigahertz frequency, it bases the frequency in which the EM signal power, use your phone upgrade, and 6G, that was probably to be the new Final Frontier wireless communication, where we have gotten much higher frequencies in the case for getting up to 100 gigahertz, I can offer benefits that we can't do now, specifically, because these high oscillations because one do like very small wavelengths, we can now be able to do send high database within not as gigabit per second, but terabit per second, we can also do it for like high sensing application that can be possible. In this case here, we will now couple both sensing and communication all at once, by using these frequencies between new applications that wouldn't be feasible or impractical to use prior. And one of them, for example, is by the use of food sensing here. So 6G signals, for example, because I'm referring to frequencies above 100 gigahertz. They have, they say, the middle ground of benefits in the sense that basically, if you go to higher frequencies with an even optical domain, like light, it can have a much higher database. Sure, but it becomes very hard to use. And so you know, why was that because it's like wireless path loss within the air. So while you do get lots of information from using optical information, you're limited in how much you can actually obtain from it, especially if you don't have close contact. But then if you go to lower figures than we do now, you might be able to have lots of wireless communication and a lot of ranges, but the amount of information you collect from it itself is very limited. So it is reduced to vague and rough cost estimates. But it's up to us because what we are working on can kind of get the best of both worlds. This has high enough frequency to get a large amount of information, but still low enough that it can be used for much more particle distances in range and basically in much more practical settings. In this case for food sensing. What we were trying to do is that you know if you go to the market, or you want to buy a banana, apple, name any food you want, basically you would see if it's ripe or not, and you do all the time metrics you can think of like for example some people take watermelon and, you know, you knock on it. Like avocados, take off the stub and you look at it. banana, you can tell by the color, but it all ends in qualitative data. There's not a real way to determine it because when something rots let's say an apple, right an apple looks almost exact near the end and becomes mush, but it looks the same. And it's like any fruit it basically rots from the inside out. So these metrics though they do work and you know, they've been working for years and they're highly correlated is very limited how much you can do on a consistent matter. And this adds up, especially on a mass global scale when you produce it yourself, because you have to assume that these fruits, for example, have been either picked or have been basically shipped out at the right time of ripeness here but you can't always tell just by knocking on some object or like, looking at this color and say, Oh, no, it's gonna be ripe. You always see a banana, that looks great. And you open it and it is all messed up. So the point is, how do we do this? Basically, how are we able to detect the food without, you know, destroying it through the process, like if you open it, okay, if you ask now you can't ship it or eat it. So this is where the sub tears frequencies come in. Basically, we have 6G signals. And it kind of gives them I was no point here that like, we want to do them as non invasive. So first, we must mean that it doesn't destroy the fruit in the process by ascertaining evaluation. But we also want to be able to get high information from it at the same time. So if we use, for example, econometrics, you can actually do this non-invasive sensing, but it's not very practical, because you have to put on close contact here. And it's very, like setting time consuming. Even then it's very hard to ask you to do this on a very low scale, when you have like, you know, fruit on a conveyor belt, like many distribution centers. But in this case here, because it's up to signals, it actually allows for larger distance, larger propagation distance measurements, we would actually now we'll say put it on a wireless platform and kind of like when a fruit goes along the conveyor belt, we just continually scan it. And you could say, Okay, why can't we do this with Wi Fi or like communication systems now? Well, they can but it's not really accurate. That information can be basically it's kind of ripe, it's kind of not ripe, same thing as me looking at it and saying okay, I think it's ripe, it's not ripe. So particularly the advantages are very limited from here.
Speaker 2 6:47
Yeah, that's so interesting. I know, like, when I go to the supermarket at home, I do like to pick off the top of the avocado to see if it's ripe. But I hadn't known that you could also test it in a quantitative way like that. So that's really interesting. I'm wondering how you thought of combining electrical and computer engineering with something that could have implications for food waste, and how you got that idea in the first place.
Speaker 1 7:06
So this was a product we dealt with in collaboration with Microsoft itself. So Microsoft itself has a huge focus on trying to actually reduce food waste itself, from the producer to consumer aspect of it. So there's a couple ways we worked with our lab here to kind of format okay, how can we use our expertise and our use of subtest signals with Microsoft itself. In one case here, we are trying to solve the problem, the magnifier service is basically how you distribute the food or how you distribute the fruit itself, as is that the main one, the main limitation itself is that because we don't know, for example, if a fruit is ripe and not in a quantitative sense, for example what most people would do is something called random sampling. So they basically just have a bunch of the, let's say, apples. And I'll test a subset of those apples for example and he might do destructive and undestructive means, but then they will assume that they say the quality of those foods is going to extend for the other. And that's not necessarily true. For example, like I always say, I have a twin brother actually. And it's a common thing to say that because I can do this, he can do the same as well, that's not necessarily a fact. So we try to figure out how we can actually reduce this problem by using the subtest accessing high sensing capabilities. And this is where the idea came from, that was why we use subtest signals to actually create this kind of quantitative measure of a given effect where we can actually purchase on a platform pretty easily. And basically do this like this, like a conveyor belt like style.
Senna Aldoubosh 8:29
That's really cool. Do you plan on doing like further projects with this technology, it can be related to produce like quality detection or just like in general so
Speaker 1 8:39
That's a Yeah, so what I'll focus on is on the producer end. Basically, how do we reduce food waste in the producers but consumers are also to blame as well. Like if you go to a store, you're gonna pick what you think is the best you're not going to pick the most unreliable one but you again don't know and no setup can work well in a warehouse or distribution center where we can have a high a high profile high sensitive system it's not a reasonable position to expect this for everyone to be able to have access to this here. So what we're working on now is trying to bring this kind of food sense food sensing on a consumer end and that can be done on like our mobile phones or some small device something that can be done on the most scalable level by getting a handheld system form.
Senna Aldoubosh 9:19
So like theoretically if like if you were to give it to consumers would it be like I would just pull up my phone and like oh wow?
Speaker 1 9:26
Just like taking a photo here and scanning the fruit itself.
Senna Aldoubosh 9:31
Oh awesome. And other than produce like what could you apply this technology to?
Speaker 1 9:36
So you can apply it if it's up to us to produce, like you can also insert into a lot of things as a medical purpose. Like skin layers can be a measurement here. So we have talked about being in a lab or sample that we have tried to get some measurements about foods. Basically it is good quality, but when we would get a quick measurement for say someone's health or something like it's going on, they can just hit critical assessments. In this case here, The same principle was carefully monitored. We do in our reserves, we kind of monitor changes in the food over time that can be kind of applied about the changes within a corresponding user. Or, as a person who needs to be checked.
Senna Aldoubosh 10:14
Wow, really cool.
Speaker 2 10:16
Yeah, I had a question about like, just to clarify. So if you would take a quantitative measurement of the food quality of some food object, then how would you? Like what would the measurement come back looking like? Like, would it be a number or?
Speaker 1 10:28
Yes, it will be a number. So generally, the general method there is many, each one has a different type of quality. What's most common is by using BRIX. BRIX is a straight cut in the fruit itself, or dry matter indicates the amount of like, non soluble materials in the food. And the main idea itself is that those things have been proven in many studies in the past to be very highly correlated to the ripeness of the food. Whereas when you hit a high Brix value, that means it's generally ripe, same with dry matter. There might be some very easy to use fruit here, but each one has been studied well in the past. So we do that based on our measurements, we correlate, we estimate these metrics fixing dry matter to then use them and we might reach by percentage. So we'll say for example, 20% or 15%. And then we correlate that itself to Fruit's ripeness.
Senna Aldoubosh 11:17
That's really, really cool. And then before we get to our last question, is there anything that you want to mention that we did not ask you or we've missed?
Speaker 1 11:26
One thing I would like to add is that one thing that also makes this part, at least to us interesting, unique itself is that we generally in our lab work in general, making system design itself. So when we made this we made it in mind, not just to focus on one fruit but multiple foods. So you try to create a model that can work for not just with say, for example, just an or only an apple, because some systems can measure really well how an apple rottens, like lots of vitamins, but also, say for example persimmons. I mean, I didn't know what permissions were until I did this project actually. Let's say for example mango, avocado, and they all have different properties, right? Because for example Apple, they have smooth surfaces as well, then we say avocados, they have a very thick surface, and they have very high thickness. In our approach, we had to actually work and say how can we always give it to you and get some kind of metric? And how can we always make sure that we're not comparing like the rightness qualities on Apple would have with like what an avocado would have. Because avocado is not as sweet as an apple or banana or like any kind of food. So how do we kind of separate them and kind of measure them on their own. So what we did is that this measurement itself took entirely about a month to do and what we did was we took multiple foods from here and created our own framework we call geometric analysis to form, this form this kind of food fruit ripeness metric.
Speaker 2 12:53
So I was actually, you seem like really passionate about this. And so I'm wondering what you enjoy most about working on this research.
Speaker 1 13:01
So I like how this is an emerging field itself. So we try to take we try to take advantage of these substantive signals, but I like is that we're able to actually create a feasible application that can actually solve a problem we've been probably getting since the beginning of time of like any type of food, farming, or food producing itself, in turn produce a potential significant improvement from here. And I like that the basics of all bases will focus most of his time on solving problems that are more applicable, like we can actually apply this to a real world setting. And then this can be out on the market.
Speaker 2 13:37
Yeah, that's really interesting. And so kind of too close, do you have any advice for maybe people who want to get involved with this field or just for consumers who are thinking about food waste, as we kind of approach Thanksgiving?
Speaker 1 13:47
My advice is what made me want to pursue this field itself. The best way is that, to do this, you need an interdisciplinary approach to it itself. So as I said, they've got group work as a team itself. So I am like the only member part of it here. And this comes from multiple different aspects, people who work for example, with foods, for example, people who work on the manufacturing, on framework, like this, distributing it, as well as just as you're trying to work on a more physical level of it, like the physics side. So when you try to pose his work, don't try to see this as like, once upon that, okay, how do you solve this as a, let's say, engineer, how do you want to say how do you solve this as a whole? How do you solve it as someone who has to ask for example, deduce the food and distribute out how do you do that? Okay, make a setup that can actually work so that they can actually apply themselves so that you can make some great equipment, but if the equipment, for example, can't actually be used for the people you're trying to help, it's kind of pointless. So you need to have this kind of collaboration to make all this work.
Senna Aldoubosh 14:50
Awesome. I really liked that piece of advice. I think like, especially like in previous episodes, we've had different engineering professors come and talk about how there's a lot of lack of interdisciplinary knowledge within engineering as a whole and like how important it is to have that context when developing technology, especially to help people with different things. So I really, really liked that you brought up that point. But thank you so much, Atsutse, for joining us today. It was really cool to learn about your research. And thanks.
Speaker 2:
Yeah, thank you Atsutse.
Senna Aldoubosh
This episode of B cubed was hosted by me and Ria Tomar. Sound engineered by me and produced under the 147th Managing Board of the Prince. From the Prince, my name is Senna. Have a great rest of your day!
Transcribed by https://otter.ai