 Yes, let's start. My name is Big Auliaq. I'm from Olufamix, one of the longers. And we are today here to tell you about our approach to Olufamix. That's the most affordable platform available currently. And because it's most affordable, you can buy it. You can have it. You can build it. And it's going to be open source. We have all these link-up links on our website. See? And that's very much it. Yeah. I'm mostly an embedded guy doing electronics. I'm a bit bored, but in this project I have been doing the right walk of talk. So, you can know that that's a much better approach. Anyway, let's start. So, the software is fully equipped. It's in version 2. And then find the link-up. And the hardware, of course, everyone is interested about how to license that. The answer is we don't know. And we are not giving too much heart with this because it's not finalized yet. We have new versions coming. It's still open all the time. We have to look in what we're going to produce. And the 3D sensor is the hottest thing we have, because nobody else has it. Okay? And we found a new way of making 3D vision from machine 360. And 150 degrees vertical. And we're going to save the demo after our presentation. And this is called for participation. Somebody's been telling that they're mostly hardware-wise. That's true. And we, the software, make this thing. Making a robot, affordable robot. Making a game. Why we did this? We did this because we needed it. We needed it for ourselves. And we actually checked different options. And the prices were from 20K to up to 60K. And they didn't even do the time we wanted. We have a small warehouse. We have about 30 customers. And we didn't find this solution. And we built this one. And I see the picture. It's summer 2016. It's a bad picture. We can find it from the web page. And it was actually remote control toy. It was working, but nobody could actually use it. Because of lagging. Network connection. And we decided we need other way on this project. And I wrote this design principles. It has to be simple. Robots must be capable of taking whatever comes to it. And manufacturing must be very easy. So we can make as many as we like. We probably need 10 or 20 of these. And since the last three documents, we allowed everything we could get from China. And it didn't work. And it has to be able to do tasks we did. And able to map surroundings. Because we don't have time to map things. And the surroundings are changing all the time. And by accident we found out this way. We found out this way that we made the platform that there's an application later. I tried to draw all sorts of different solutions on this system. It was like a tower. And you see those poles. Robotmakers made. We tried to make something like that. And it didn't work. So I decided that we make first the platform. And then we designed what we do with that. And that's actually something that we found out to be a good solution. The platform is the product. The integration is simplification. It has to be very simple. And because everything is integrated, it's very cheap. Very cheap in many countries. And this is from Pikipalia. The definition of algorithm is wrong. It's mostly the same thing I figure out myself. It has to be able to gain information about the environment. Learn. Work without human intervention. And move through the space. And not harm anyone. Now comes the big thing. The breaking news. I'm going to use Ross. But we are going to support it. And we are going to need some help for that. And maybe I'll tell the reason why. Because we know that what others might be doing using Ross. And we know that with Pikipalia they buy a lot of these expensive modules and build the robot using these blocks. And it's quite easy. You can just go and buy these off the subcomponents. Then you can do the robot. But the robot comes here. These extensions are up. And then you have a deal of materials from 5,000 to 10,000 K. And then you need to try to make a product. And you have to. So we use it on top of it. And it's going to be that big ball of 30K of Ross. And then we use it in the background. Then we require you to use that set of less. The game is between all these modules. And you are going to convert all kind of sensor data through USB pass. And then you need this message passing and all kind of stuff. And it's nice when you want to do a product for your life. I think there has to be a better way. And being an enterprise designer, I just think of all of it in a single board. And this great thing is that it gets so much cheaper. And then all of those sensors copies is done on this SDM32 microphone. And, well, it's got a few bucks. And it can process all the data. And it's very low latency. So we have to center data. It's very short. It's very quick. It's great. And it's cheap. And it's very slow to maintain. Then we... I make the first version of this board in about six weeks. And this is our first batch. And then we... This is the second version. This is that actual product produced in China at least. This one. But then we went to the land. This is a simultaneous localization of mapping. And it's a very big question in mobile robotics. And it has been analyzed academically for decades. And it's probably likely because if we want to localize, we need to get the coordinate of the robot. And we cannot use GBS because it doesn't provide doors. And even out of the resolution, very poor environment. But we need to know where we are. But we don't have the map. And to be on the map, we need to know where we are. So we just start building the map and all the errors will appear. So this is a difficult problem. The problem is, there hasn't been much open source involvement in solving this. For example, with Ross, everyone is still using G-Mapping with this demonstration algorithm. We got by... we started this 10 years ago. And it's a great algorithm, but we don't... So we need to learn how to do this. And right now, we're using any special, but it's usable and we are going to develop it further. And that's great because we have all the sensor data. We have problems with our hardware. We can't change our hardware. We are not open to any other hardware. So this is an example that after about... I think it was four months after we started from scratch and I had to go for a 10-minute and about two weeks at that point. And there is no route going on here. So the problem automatically is for my parents' house in this picture. And well, after it has done this automatically, we can move around and go there and decide the best route. And if you try to do that, it will go between and it will wheel route and things like that. And we have another example from yesterday. We went into that journey building and that's the map that you got, but it's all in there. So with this actual machine and people thought this is some kind of lost car. They didn't even realize this is a road but they tried to stop it and move it. And they didn't try to move it but then it's powered. They tried to keep it in position and people are very surprised. Yes. So we are stunned. Did we do something new? Find out we went to South Korea. Overgrown 2007 fare. 2017. In September last year. And we just wanted to know is there anyone doing the same thing and maybe we probably didn't find everything from Google. And what we found out is that nobody else knows how to do the thing. And people were actually lining up on our booth and they were complaining after two days that they hadn't been able to talk to us. And then we found out we have done something unique. So maybe we should keep this out as open source. And that's what we do. Yeah. So our next steps the basic hardware is really good I think. We are going to do one more modification to make it even better but no big change there. And then we tell the company it's really simple. I'm going to talk about it in a few minutes. We are making small batches. We make 30 of these. This is one of the 30 from our project. We are making batches bigger and bigger and we are making small changes. And we are open sourcing all software, every single one software, and as we said all software is unclear, it's clear that we will give all the schematics and everything like that out but we don't decide it about the actual cat parts yet. And we need application development because our great idea is going to be called the platform model. So our platform is the smallest component instead of a sensor. Our component is two platforms and that makes it quite easy to develop your own application. This is the delivery boy application here. It's a demo of the simple possible application, the passive application right now. I repeat this will have a few cameras and I think it will be more usable. To make it affordable we have to take the price down so that it costs about 1,200 Euro to build this. So we are selling it out at 3K right now. We need to have some money to do this rolling. Yeah. And then we are going to do the smaller one, the S model and for that we are going to do a developer special where we are not going to make any profit basically. Basically the 999 offer is something that we are not sure that we actually make money at all. Yeah, because we need this to go out for people to find this and start building the ecosystem. The software architecture I know many people here are interested in the software. It's right now done by me with 90% in C with 5,000 lines and we are out of it. We are currently in SCM32 microcontrollers and no any kind of error is left. Well, very few dependencies and that's for a good reason because I only use what I need. This is on our website just an example of the project that we will find on that paper because we are there and then we have the Raspberry Pi inside we are not going to do that we are not going to slam and we won't find any complications that's one of the problems. And then being a prototype we are going to replace them with properly done versions. Let's get back to where we were earlier because this conversation is a little bit unthero because you see the real error of this work that the laser scanner that's hugely expensive part there are some cheaper options but that's often used in robots so we are we need our own sensor and then that's how we imagine our prototype in 3D done with light sensor system and we are going to actually a lot of tasks and we are rotating through the dimensional what makes of playing and it's not good enough because as you can imagine when we have table of this profit then you cannot see you can only see at this 2D level and that's why we need sensor and really time of life that new it's a decade of technology so we take this normal laser scanner which takes one pixel of the volume it measures the distance value instead of a single pixel we have a full sensor with 10,000 pixels so it can produce a problem like this or Iran shape so it produces points out directly and that means from our current 5D 2D example per second we are going to get half a million the distance measurement per second in 3D of this and we are going to bring all of that sensor with this one single vector it was done and this 3D cameras have been available to self for quite some time but they are neither fat or they are really expensive and they are going to pretty much cost down so I'm going to give you a little example hopefully works so let's say the cables are connected on so then the distances are actually all and probably in that image you can see for example I'm freeing right now and when I go further back that it will be much better so you can for example see your own case in that picture and then we can measure how many meters it says that one I just moved there and now the game is going down to one and one what is it, how many meters so yeah, that's kind of technology here and this technology has some drawbacks it suffers from some issues like grey light, if I put my hand here I think that background being just there so we talk probably for a while better on our website because we have some unique ideas how to fix these issues and instead of that we are going to get them out freely to prevent all that fat controlling we don't want to be a part of that so if you want to see a better demo come to us afterwards we will show you so yeah, this is basically our sensor, it's a little big right now but this is being manufactured right now in China we also get them before the Chinese New Year it's a smaller version yeah, I have to fill up here it's like the current sensor would be like 40 times 40 millimeters and it's much smaller this is the very first one we ever made and in the new year we are lining up with the manufacturer's and hoping that we don't get it before the 16th this month and if we get it we are able to produce more robots the first ones but I'm actually didn't feel that we have 10 of these sensors around the machine providing 30 seconds because we can do this sensor for about so we can put 10 of them and we have 300 at that point and it's still cheap and we can see that the image on the back would be placed on that we get full 360 degrees by 120 for 110 to decrease that way see the white paper on the white page yes, let's make all these robots affordable again these are the gross solutions everyone is telling me it's $10,000 make one more to speak whatever you can do it's good and the price I have told you is very cheap now we have to make any money by selling that price but what we want to offer you it's a handout you get the technology but unfortunately we are unable to deliver before June and that is 4 months away right about on the first thing at our web page there's a link where you can make your order and there's also a developer survey if you want to participate you can leave your email and we'll contact you