All right. So I use this photo for two reasons. One, this was my my better years in and out. It was like ten years ago. And secondly, this what this photo was on Louisiana State University front 84 months. And the text says, Louisiana State University students doing all spill surveys. All right. Okay. I was a professor here in case you didn't get it. Okay. So why are we said Louisiana? I was leading a team of graduate students and undergraduate students to actually do our spill survey there. We actually developed some of the early versions of economists surface vehicles in and then deployed it. And one of the title that goes there to try to find remanence oil contents. So that's why I was sitting there trying to not sitting there sitting in a pedal boat and trying to help the vehicle back. It's having problems. Okay. So that set the stage of what I do. And, and here are some of the data we collected at the lagoon. So on the left you see the bottom shape of the lagoon. On the right, you see some kind of oil concentrations that collected by our vehicles. Deploying robots in these kinds of environments are tough because it's very hot. And it was like a hot Louisiana summer near a water. Lots of mosquitoes like SAS can maybe someday viewed robotic mosquitoes. So anyway, so then we think, how can we do this autonomously in the future? So one of the things I had been doing over the last ten years is to study fish, right? And draw inspirations of fish. And think about how we can imitate their swarming behaviors. I don't raise fish in my lab, but I collaborate with one of the best in the world. The name is in cousin at, in 2009 he started as assistant professor in Princeton and Naomi later and I working with him. And he actually have this very at that time state of the art test back to that can track the movement of each individual fish while they were doing swarming. And also, we actually put robotic fish inside this tank together with the real fish and test their reactions. Remember this was 2009. It was pretty cool at that time. And then from that, we will learn a lot of inspirations and we translate the fish, the insights on how fish does it to mobile robots on the left. So you can see that the robots at the beginning sort of work individually and cannot go anywhere once they decide to collaborate, effectively, track the gradient and go to the source. And later, we will generalize this to more sophisticated, sophisticated behaviors where the robots can also avoid predators. All right? So in, in the real world, actually, source seeking is not that easy. Oftentimes you don't have a well-defined spatial gradient. So you oftentimes has kind of plume structure underwater. So by collaborating with to Georgia Tech Professors in about department don Webster. Yeah, that's the thing you want to. So so so we study graphs and then we actually translate this, the flume cracking behavior into, into robots. So, so here is the crab inspired plume source tracking behavior. So this is more realistic. On the left-hand side, I'm showing a realistic underwater plume near sea bottom where a lot of minerals can be found. Okay. So, and then how can we make these things work in the real ocean? Over the last about 20, 25 years. A kind of robots. Robots was developed and really changed the face of Physical Oceanography. So it is called underwater gliders. So what it does is it actually move in the ocean without consumed much energy, does the changes buoyancy, and then leveraging the fluid dynamics to create this horizontal motion. Okay, So it's sort of swim in a, in a kind of zigzag shape, shape and the water. And it doesn't communicate under water. When it needs to transmit data, goes to sea surface and use a GPS to obtain its location and also transmit data back. Why does this change the game? In ocean observation? Previously in ocean observation, you have to use men. You know. Labor intensive method. You have to deploy a boat and all the humans. A lot of humans are involved with ships. It's very expensive. And this is, this robot is low cost and also its long endurance. They can stay in there for very long time, several months or now that it's even years. Okay? And it can also harvesting energy from the environment. Lots and lots of recent developments. Host loads of sensors. And as long as you can make these sensors small enough and low-power enough to be put on a robot and it can host it. But that's completely changed the game from non-autonomous, semi-autonomous sampling to autonomous sampling in the ocean. So, so, but once it comes to autonomy, autonomous sampling, starting in 20, in 2006, we start to develop softwares to make the autonomy into reality because this thing works 24, 7, it's like a Mars mission, right? So no one can, no human being can go, it can sit there and control it every day. So we develop this autonomous software that takes the ocean models right and get the data into the ocean models, make real-time prediction for the ocean states and do path planning. And then make the robot move in the ocean. It doesn't sound very exciting now, right? And because of the autonomous driving, but this was 2006. There was no autonomous driving at that time. Okay. It was autonomous driving in the ocean. And Q this states were still doing this. So these are some of the recent deployment we have and we keep overcoming challenges in the ocean. For example, we have really, really fast flow sometimes that much faster than the gliders the end. And sometimes we have these fast-changing see conditions over an area where the robot has to move from one side of the front to the other side, the front. Okay. And also sometimes we have friends in the ocean that thinks the gliders are there. Mate. So anyway, so we have to overcome all of these environmental uncertainty. And this is at the very core of robotics. And also, of course, why don't we use a lot of them because they're much cheaper than, than, than ships. So this was also in 2007, we develop some coordinate controller, of course, in collaboration with Princeton and to coordinate up to six gliders at the same time in a real realistic ocean environment. So this is real, this is data from the ocean, okay? And, and things might talk to you about maritime robotic network. You may wonder what it looks like. This was our vision 2006. Okay, and you can see this, this is pretty cool. Even looking from nowadays. In other words, all the different assets try to coordinate their motion. Try to autonomously, right? Getting lots of swarms right together and exchanging data. So that we can automatically collect data from the ocean instead of relying on human labor. Okay, this was 2006, and we're pretty close to it as we're speaking now. So along the way, in 2012, the idea of distributed network comes into play. So the idea is, would be, why, why don't we have the robots actually share information while they are in the ocean with each other instead of transmitting everything back, right? One of the biggest challenge here is underwater acoustic communication, because in the ocean, when you tried to talk to each other, probably acoustic is the only way when, when it's beyond 100 meters. Okay. So, um, but, but underwater Acoustic, it is pretty difficult. This is, this is how sound propagates underwater. Lots of Marty pass, lot of complicated physics you have to think about. So we start getting into this. Okay, up until now I would say I know what a network is about 2D and above water, we're talking about 60 now. So it's just huge gap there that need to be fail. So once, once you get your self into what are acoustics, you open a completely new world. So for example, you could use acoustic tags to track fish. So these tags are really tiny. Fish swallow them. And when fish swim in the ocean, right? When they pass through the receiver, you know, hey, there's a fish, right? So this actually is the data that decides whether you can fish in certain areas or not. And let me tell you, very accurate, very, very scientific. I don't think anybody can buy jokes. That's weird. Yeah, I think so. Yes. So so so basically I'm trying to say it's very inaccurate, It's very unscientific because, because it's very reliable, you know, it's, it's really difficult to gather these data. Well, why not use Roboto, track them. Right? So, so that's, that's, that's this project is about. And also what, why don't we deploy swarms of robots to observe three hurricanes at the same time in the ocean while they were forming. So that we can really predict the hurricane paths better. Accurately. Yeah. So this is another project that I was involved in 2018. I'm not leading it that I was involved. And also we got a recent project. Ai gets into this field to, so we use AI to help better planning our passive robots. We're doing this fish surveys, so this is completely simulation, okay, this is not real. And we're still working on this. So oftentimes, I have a problem working with students and attracting students because students think meridional of us is hard. Why? Because I need to go to a boat, I get sea sick, I cannot get my paper done, right. I spent a day there doing nothing, right? So actually also a lot of you are thinking, I'm doing such so cool thing in AI. Just for everything. And I don't want to focus on the water. I just want to try whether my algorithm will work on water. Well, here's what we're doing. We know such problems exist, right? So that's why we decided to build open-source, really low cost, easy to put together devices to do, to help people, right? So this is starting in 2017, we start building Georgia terminator underwater robots. And we're still doing it, okay, and it's getting better that it's very easy to put together and you can just testings in swimming pools. Okay. And also surface well, yeah. This was 2018. We also improving on this. Yeah. And this is this boat is doing like efficient at inspection. Okay. I think I'm running out of time. Two more minutes. Okay. Yeah. So, so you can, you can actually pass your wonderful deep learning based facial detection algorithm, efficient at whole detection of it. Back in this, yeah. Okay. Yeah. And also, you don't even have to touch water to work on underwater robots nowadays, right? Because we have this flying underwater robot, which is autonomous Blimp, right? And it has very similar fluid dynamics with underwater robots. And we can, It's very easy to put a swarm of these things together. Yeah, And you can, you can do a lot of cool things okay. With this without actually touching water. Okay? And for example, we're using the blend to map the wind field here, which imitates underwater robots marrying the flow field were, and on the other hand, this is using our open source underwater vehicle to measure underwater acoustic field in a swimming pool, which is actually more difficult than doing it in the ocean. Because Sweden, who has a lot of vibrations. Vibrations, okay, for acoustic propagation, it was pretty challenging to. So what I'm trying to say is the field of Marie robotics is completely different now. Okay. It's not like you, you can now get anything done in three days just because you get on a boat, right? Just to just to collect some data. It's not like that anymore. So our effort did it went unnoticed. So we recently received an NSF funded CCI project, which is a big deal. So this is with university alabama. Put all these things together. Isvs, acoustic modems are open source, indoor, outdoor. So that we can open up these testbed for the community to get more people like you to get involved in robotics. That's in my talk. Thank you. Any questions for water? How, what, how how does yeah, there's some slightly different slight differences but mostly they're there. They're very similar. Yeah. Yeah. The one behind you, sir. Yeah. What are you going to come competitive? What has it? Oh, yeah. Yeah. Yeah. It has been there for forever. I know it started in about 2007. Yeah. Yeah. Yeah. No. Because this is purely a student effort. Yeah. So but but there's a long history. We used to get involved with one of the teams. Yes. Yeah. Yeah. Okay. I think you have someone has a question here. Okay. All right. If not, we'll thank you. Yep.