 I've been working in the LTC, it's an ASTAR research institute for the last three years. I've been doing applications using ABD robots, KUKA robots, Final Robots, UR10 robots, KUKA IWA robots. All of these are dedicated for polishing machine tending and all related activities for the application. I want to present you today what I've done using Roles Industrial and why I believe it is a game changer in the industry. I want to add on a little bit more the definition for Roles Site. I want to include you guys what is a meta package, basically a massive package, including definitions for KUKA robots, Final Robots, ABD Robots, Motoman Robots, etc., etc. So the idea of this package is to enable users to take advantage of these robots that are probably in your company doing nothing. You can go sit down, take probably 45 minutes to do a pseudo update and you get the package to install it. You continue your development just by hard coding or remembering the points that you want to apply for example for polishing. You want to do a trajectory, you define, you move your robot, save the point, go to the other point, save it and you got it. You go to the C++ and you perform a whole route, basically that's what it takes. You can create a trajectory repeatable that is good enough for you to develop any result that you want. So now you don't know the language card for example, you don't know any KIL, you don't know rapid code and you know how to use a Linux computer. That's interesting right? Only one language allows you to use all these kind of programming languages. So personally I'm not good at farming, but I managed to do some development using my robot, which is my hardware. I'll show you later what I did. It's just a funny debugging tool, it's also doing polishing and there you go. That's the main capability I wanted to show you. I know it's pretty straightforward, but I want you to change your mentality. I don't want you to believe that ROS is only for voice. ROS gives you the ability and the tools to develop your career and put it that way. So what is really ROS I had? You need to understand the overall concept and the capabilities that need to be defined to reach a certain maturity in the industrial level. We know all the development done in the academia for example, for the motion and planning trajectory. We also understand the perception and the sensor fusion that the work that it takes to do our sensor fusion for example. All these abilities are just simply a click away from you and you can just simply download the package, put a couple sensors, one Kinect and then there you go. You have a defined environment using our base and you can perform any application. So what is a real headache? We need to define that and you define that. The industrial environment defines that. So let's say you have a customer in Siren and willing to perform polishing again and improving the capabilities of their components. Let's say polishing iPhone cases. You have a solution now. You just have to adopt whatever you know in ROS, whatever you learn in university. Provide this solution to your boss and then you have a class. So ARTC is a proud part of the global consortium which is America, Europe and Asia. And ARTC is part of the Asia side. We have SWRI in the US and we have Frank Holtler in Europe, in Germany. All of them contribute constantly, every day, providing packages for you basically to play. Take the robot, see any problem with any topic that you have in your code and then you continue your development. So it's constant support from these three research institutes. So you're not alone. You have people helping you, backing you up. I want to highlight on the next event which is a week ahead. We're going to have keynote speakers. One of them. He will be sharing all his knowledge. You guys are invited to come across, ask questions and get to see what are the opportunities that you can take from here. So as I mentioned before, what is it capable of? Okay, we have officially distributed, this is the software that is backed up by the research institutes. We have ABB, FANUC, MotorMan, Universal Robots. What you can do basically, what Juan was doing with this robot, just take the topic, the sensor, MSDF, the messages and instead of reading the CMD velocity, just read the joint trajectory, the joint states. Once you read the joint states, you can basically identify the position of the robot and then you have constant real time feedback. I'm going to say real time in this case, but you can see how your robot is behaving at that particular moment. So you have a visual feedback. You can identify your TCP. You can define which trajectory you want to follow. You can use packages as the cut. If you have a really complicated trajectory, you can use plugins included in the artist. Cartesian plugins, for example, which my colleague will present later. You have everything, you have all the tools here. And these kind of options outside are very expensive. OEMs will tell you, you have to pay for this module and then you can do this kind of welding, for example. So now you have the opportunity here to take the robot, perform a tool path, define a tool path, generate it yourself and perform itself. That's what matters. Now you're saving money here as well. Furthermore, you have the capacity of including different sensors. So in the integration site, probably as you saw before, it was using lighter sensors. In our case, we can also include, for example, robotic grippers. We just have to download a single package and from the TCP communication, for example, we got a sensor and we're ready to go. Same case, for example, for the Intel RealSense. Intel has compromised to develop their own packages. They showed it in RustCon last year. They had a nice package for the RealSense. You have four cameras, you have depth, you have 3D tools, basically. Not only Intel, you have a lot of companies compromising to develop their own packages. For example, SICK. SICK is constantly helping you for 2D sensors and 3D later sensors. For safety, most importantly, you can also include safety sensors in these cables. Then you have real-time tools for integrating communication as EtherCAD, EtherCAD buses. You can just simply download the open source real-time EtherCAD and then you integrate it into the package. There you go, you have iOS. You can now communicate with the world-activated actuators, activate valves, receive signals from different type of environments, etc. And more importantly, we also want to have cheap devices that can communicate with the world. So we have raw cereal, right? We all learn nice things at university, how to use arduinos or raspberries. So what is the best way to communicate from my heavy industrial robot to kind of cheap environments? We can use raw cereal packages, same as what it did probably in its development. Just one package, connect to USB, and you're reading topics from the pie or from the arduino as well. So I want to show you one process we developed in the company. This one is dry ice blasting. It's basically for cleaning heavy components in the aerospace industry. And I'm going to show you now. So again, we decided we want to perform all the world's launch thing. Try to go ahead. The environment was developed by Inten at the time and he performed all the measurements and he accommodated the component as well. And at the end of the day, he just created a trajectory that was in here for cleaning the component. So I'll present you also a video, parallel, and remove the environment so you can see better. So that's the real application. So you see probably a large robot have a ton of payload and you can still control the roast. That's an amazing tool. We're just spraying particles inserting kinematics in every particle. So once it reaches, the component cleans the surface evenly without damaging it. So these are the kind of applications that RoSci do that. So now you don't have to think only about creating small toys. You can use big machines safely and accurately from whatever task you need to do. Without knowing, for example, Rapi, this is one of them. I'll present you now the polishing. This is a polishing for aerospace components. This is a bit treated here. So we have a nightmare itself for doing debarring in the company. And a lot of people struggle to program this robot. And I don't believe that. It's very hard. We all know little of Carol's programming but we took RoSci to do this and it turned out to be really easy. What we do here is a polishing of aerospace components. I kind of showed you the aerospace component. I'm just showing you a dummy component. But what we do is basically define the trajectory and then perform a motion, not linear motion at the moment, but it's just doing a trajectory definition to make it easy to move it. We also do a collision avoidance. We define the whole cell and then in side-arbies we can see the trajectory planning without any collision. And we do this multiple times for all the components. So wrapping up, I'll try to be honest here there is no perfect software solution. That means RoSci is not the only option that we're going to have. But it's a very good option. It's a very important option for you to use in case you have problems, in case you have difficulty to use any other kind of software. It's a solution for you. There are different features that are missing, like definitions of speed, for example, controlling the definition from one point and then changing that to the other one. But at the moment, you kind of make a similar statement with ABD programming, for example. Then you're going to have to have a C++ or Python, but I guess we want to spy here, so it's okay. Then just to finalize, just the real opportunity here is that the interoperability between the users and the multiple robots is increased. If you just have the option of learning RoSci, you can do any robot. That's it. That's simple. So it's an opportunity for you. Thank you so much for listening to me. Just remember to come to our event next week. Is there any questions? How do you relate to RoSci with the other robotic computer? Did you have regime software or did you just connect the TCP and then define the IP address of your robot and then you just bring up, I think you saw the RoS launch name of the package, bring up then the robot IP address then simulation folks and then he will come up and then you have access to the joint states of the robot, for example. So that has an info on so you need to have an integrated robot So in the case of final you need to install two additional items. Or in the case of ABB, you need to have phishing in the face, for example. These are the current, as I mentioned before these are the current officially supported. Then you can find packages for KUKA which are not totally supported you might find but it is known that we know that they divide out this kind of function. There are some nodes on the disk. So that means that the manufacturer has to provide that kind of... But for example, in the case of final, you have the position streaming, right? But that's because of the ability of the controller not because we don't want to develop it. And then this is a small advantage, but you can see the real time during state then in the case of the UR10, that's saying you don't have it, then you have more options. At the end of the day, all of them can do the task. That's what matters. These are little details that can be developed with the collaboration of the OEMs. That's kind of what you're looking at. And to add up, I think the packages are there. For example, some of the not-supported Alps, usually they do a wrapper based on the existing APIs of the arm, and then they just convert it into a ROS package which is something that ROS can understand. So they are official wrappers of the party doing programmers. But the only thing is that maybe you don't have a full support from the member. And in some cases you're not allowed to use the visual definition of the robot because of legal aspects. Is there a recommended player who can find this kind of kind of... He's a nice person for that. Yeah. One player or is it actually distributed over different applications that you can find for the particular robot? No, no, you just go to the ROS-I website and they provide even a tutorial on how to start developing. So you can download, develop in tools, you go download the ROS-I package depending the ROS version that you have, the kinetic, you download that specific package, the meta package of the ROS-I and then you're ready to go. But only for the official ones. And then you can also go inside the ROS-I and then find for FUKA for example. And you can do in-plane. You can explore all these activities offline. You don't need a robot as well. Any more questions? Okay. Thank you very much for this base and I hope you enjoy it. Thank you.