WEBVTT 00:00.000 --> 00:16.320 Diorium. You're amazing. Last talk of today in this room. And that's full. Great. 00:16.320 --> 00:25.320 I have been three guys, but check out an old VW Perlo. 00:25.320 --> 00:34.320 That's great. Mark, Loiq, Tibu, and converted that one to an EV. 00:34.320 --> 00:40.320 As I'm myself, EV driver, I'm much impressed what you did. 00:40.320 --> 00:46.320 Well, they haven't seen the slides yet, so maybe it's not that impressive. I don't know. 00:46.320 --> 00:47.320 Have fun. Thanks. 00:47.320 --> 00:54.320 All right. Thanks. Hello, everyone. 00:54.320 --> 01:01.320 So indeed, as mentioned, this is the team. We founded the company a while ago. 01:01.320 --> 01:06.320 We were lucky enough to found a startup that also got acquired a few years back. 01:06.320 --> 01:10.320 And when we left the company that acquired us, we wanted to try something completely different. 01:10.320 --> 01:14.320 We've been working in finance, building open banking aggregators. 01:14.320 --> 01:18.320 And so we really wanted to learn something completely new. 01:18.320 --> 01:28.320 And we took all of 2024 as a sabbatical to just play with cars, like what can three boys actually do, but play with cars. 01:29.320 --> 01:32.320 Why did we actually decide to tackle on that challenge? 01:32.320 --> 01:38.320 Because we saw that the transport industry, automotive, but also when you look at all the types of transportation, 01:38.320 --> 01:41.320 can be not a cool transportation, whatever. 01:41.320 --> 01:45.320 It's kind of siloed and also a bit closed. 01:45.320 --> 01:51.320 It's not like that much open source, big open source projects in the transport industry. 01:52.320 --> 01:59.320 And we also were seeing more and more scary stories about cars being hacked or EV charging stations being hacked. 01:59.320 --> 02:08.320 And all that stuff, and we were like, okay, but what is the actual lifespan of software in cars or any kind of other transportation? 02:08.320 --> 02:10.320 And what does it actually mean? 02:10.320 --> 02:19.320 Because we feel that maybe that just a feeling again, that the security and also the safety of all that software is a bit opaque, 02:19.320 --> 02:22.320 very specific to each manufacturer. 02:22.320 --> 02:27.320 It's not like what we have in the open source where everything is kind of transparent and open. 02:27.320 --> 02:31.320 And there's no real off-the-market software for cars, right? 02:31.320 --> 02:37.320 It's not like you can just buy a used EV and install the Linux of EVs on them. 02:37.320 --> 02:41.320 So what does it actually mean to replace the brains of a car? 02:41.320 --> 02:46.320 What does it actually mean to create that new brains software brains of a car? 02:46.320 --> 02:56.320 That's a little bit how we got into this hacking project, which is going to look a lot more like a hobby project than a real product, but anyway, come back to this. 02:56.320 --> 03:03.320 We also noticed that it's very difficult to make parts from different manufacturers to work together. 03:03.320 --> 03:09.320 It's not like you can take a breaking system from Mercedes, for instance, and put it in a VW. 03:09.320 --> 03:13.320 They don't talk the same language, so how do you tackle that? 03:13.320 --> 03:18.320 Because in the end, you have plenty of components that basically do the same primary function. 03:18.320 --> 03:19.320 It's just breaking, right? 03:19.320 --> 03:23.320 So how hard can it be to make them work together in the same way? 03:23.320 --> 03:28.320 So why do we have to have a completely different languages that these components speak? 03:28.320 --> 03:31.320 How can you make them work together? 03:32.320 --> 03:37.320 So we started thinking about, okay, we're basically talking about a grading vehicles. 03:37.320 --> 03:51.320 Like, it doesn't have to be just converting it or whatever to an EV, but it's also, how do you make sure that your car is upgradable or that you have an aftermarket solution for your car or any other vehicle? 03:51.320 --> 03:58.320 It's much like you have opportunities to install Linux on your old computers or post-market OS on your phone or what not. 03:58.320 --> 04:14.320 So this all sorts of open-source solutions for these types of hardware, but there's no option for the new smart phones and wheels, basically, because the cars are becoming more and more smart phones. 04:15.320 --> 04:22.320 So what we mean by upgrading is changing the engine to something that's more environmental-friendly, could be engine swap. 04:22.320 --> 04:31.320 It doesn't have to be electric, but it can just be something that's polluting less, I would say, or complete EV retrofit like what we did. 04:31.320 --> 04:37.320 But also adding some basic things to existing cars like an infotainment system. 04:37.320 --> 04:46.320 Why do you have to rely on some, again, proprietary infotainment system, why don't you build your own with open-source tools? 04:46.320 --> 04:52.320 Assisted driving, autonomous driving, also something that would be nice to add on cars. 04:52.320 --> 04:56.320 This open-source project is to do this, by the way, especially the autonomous part. 04:56.320 --> 05:01.320 We might have heard of, come out of AI, for instance, or even remote control. 05:01.320 --> 05:06.320 We dig into the topic just a real small disclaimer. 05:06.320 --> 05:09.320 None of this is road certified. 05:09.320 --> 05:11.320 We know it. 05:11.320 --> 05:14.320 It's not our goal today. 05:14.320 --> 05:17.320 So we did this on our own funds. 05:17.320 --> 05:20.320 So it means we use components we could easily find. 05:20.320 --> 05:25.320 Also with the second goal that we wanted everybody to be able to get started as well. 05:25.320 --> 05:32.320 So this is not running on automotive grade hardware and all that crap, because we simply don't have the money for it. 05:32.320 --> 05:38.320 But we could definitely adapt it if the need would be there. 05:38.320 --> 05:47.320 So now that this disclaimer is done, why don't we try to upgrade this to 2007 Polo using only open-source software? 05:47.320 --> 05:49.320 This is the donor car. 05:49.320 --> 05:53.320 It has kind of a history because it was my first car. 05:53.320 --> 05:56.320 So this is the first car that I drove. 05:56.320 --> 05:58.320 And I sold it to Luick. 05:58.320 --> 06:03.320 It was sitting right there. My associate a few years after that. 06:03.320 --> 06:05.320 And at some point you wanted to get rid of it. 06:05.320 --> 06:09.320 And it got really the good timing for us to work on that project. 06:09.320 --> 06:13.320 So we said, why don't we buy it back through the company and actually do something with it? 06:13.320 --> 06:15.320 So we worked on it. 06:15.320 --> 06:18.320 And just this is a pretty heavy slide. 06:18.320 --> 06:20.320 I'm not going to read everything. 06:20.320 --> 06:23.320 You can read so you can read by yourself. 06:23.320 --> 06:28.320 I just want to show you almost everything we've been doing on that car. 06:28.320 --> 06:38.320 Also showcase the mechanical work that unfortunately I won't have time to go to much into details in this presentation. 06:38.320 --> 06:40.320 And also the hardware and software part. 06:40.320 --> 06:42.320 I mean it's only 40 minutes. 06:42.320 --> 06:44.320 I cannot go in the details of everything. 06:44.320 --> 06:46.320 So we're going to fly over a few things. 06:46.320 --> 06:50.320 Explaining how we used opens our software with this. 06:50.320 --> 06:57.320 And in order to avoid you from falling asleep, I'm going to show you already now what it looks like. 06:57.320 --> 07:01.320 And so you can see we just unloaded the car from the trailer. 07:01.320 --> 07:04.320 It was about 10 days ago. 07:04.320 --> 07:07.320 We moved the car from the place we actually worked on. 07:07.320 --> 07:13.320 And we moved it to a storage facility that's closer to the testing track where we actually going to be driving it. 07:13.320 --> 07:16.320 Because again, not to what's certified. 07:16.320 --> 07:20.320 It's like you can just drive around with it in the streets. 07:20.320 --> 07:24.320 So the car is actually as you can see, driveable. 07:24.320 --> 07:29.320 Now the reason there's a very big cable is because there's only part of the battery in it. 07:29.320 --> 07:32.320 And also it was raining a lot that day. 07:32.320 --> 07:37.320 So we didn't want to fry in the car because we removed so much stuff that water leaks in a bit. 07:37.320 --> 07:40.320 But anyway, so the car looks like this. 07:40.320 --> 07:44.320 It's electric, there's a cable, right? 07:44.320 --> 07:47.320 So yeah. 07:47.320 --> 07:52.320 And there's actually a huge ACDC converter that Luis built as well. 07:52.320 --> 08:02.320 That is sitting on the passenger seats that converts basically just standard grid electricity to what we need for the car. 08:02.320 --> 08:07.320 And so we just plugged it on the walk plug and then we started driving it. 08:07.320 --> 08:10.320 So that's how we tested the car to begin with. 08:10.320 --> 08:13.320 So this is what it looks like now. 08:13.320 --> 08:14.320 Pretty cool, right? 08:14.320 --> 08:17.320 And all that thanks to open source projects. 08:17.320 --> 08:25.320 And for some of them probably unusual technologies, things you would not necessarily expect to see in a car. 08:25.320 --> 08:34.320 And that includes a mixer, which is a language working running on the early virtual machine, aka the beam. 08:35.320 --> 08:41.320 It's also using nerves, which also you probably may have heard of or not. 08:41.320 --> 08:50.320 But that is a tool that I'm going to talk a bit more into details afterwards, which allows you to package your firmware for the ECUs that we actually built. 08:50.320 --> 09:00.320 There's going to be some VUGS, some Flutter and some Phoenix, which is a way for us to build APIs and web interfaces for the components we have. 09:00.320 --> 09:06.320 So while you're mixer, it's a dynamic functional language running on the beam, already mentioned that. 09:06.320 --> 09:14.320 The beam was really made to run highly scalable and highly available systems. 09:14.320 --> 09:19.320 It was basically built by Ericsson for telephony purposes. 09:19.320 --> 09:25.320 So it's not like it's just going to fall on you anytime soon. 09:25.320 --> 09:30.320 So it works quite well, it has decent footprint for what we want to do. 09:30.320 --> 09:45.320 And what is really great also about the mixer itself is that it uses pattern matching as a parading, which is really good when you want to parse a series of bytes or bits or whatever. 09:45.320 --> 09:54.320 Or if you want to parse messages, which is exactly what almost 80% of this project is actually for. 09:54.320 --> 10:03.320 So if you see the example on the right, that's the only thing you need to actually parse meaningful information into a series of bytes. 10:03.320 --> 10:09.320 And it really shows you how powerful pattern matching can be in those types of use cases. 10:09.320 --> 10:14.320 Another tool we've been using is called nerfs, so this is also based on Elixir. 10:15.320 --> 10:19.320 And it allows us to create a firmware based on build route. 10:19.320 --> 10:27.320 So it leverages all the power of build route, meaning that if you're not happy with the basic systems that they provide, you can still bid your own customize your own. 10:27.320 --> 10:39.320 And what it does is that it packages your elixir code into something you can directly flash on an SD card and then you can use them on Raspberry Pi or any other boards. 10:39.320 --> 10:45.320 It also supports redundancy of route partition over the air update. 10:45.320 --> 10:51.320 So it really has all this flexibility that we can definitely use in a project like ours. 10:51.320 --> 11:01.320 All right, so now that I've settled the idea of what we've been doing, let's look into more details about what we did. 11:01.320 --> 11:04.320 First thing is understanding the cars language. 11:04.320 --> 11:08.320 Like I said, they speak something together all the components in a car. 11:08.320 --> 11:18.320 And I don't know how knowledgeable you are about, or somebody even all of that, so I'm just going to explain it to everyone, but they're all wires in your car. 11:18.320 --> 11:24.320 And some of them are used for components to communicate together through what we call a canvas. 11:24.320 --> 11:31.320 So it's basically a twisted pair of copper that is running through your car on which all components are connected. 11:31.320 --> 11:42.320 And they will emit messages at a specific frequency, and then we'll also listen to messages that come at a specific frequency and do something with this information. 11:42.320 --> 11:47.320 Right, so that's in a very high level nutshell what's happening. 11:47.320 --> 11:49.320 And it looks like this. 11:49.320 --> 11:59.320 So when you hook up a can sniffer, so it basically something you can plug over USB, and then you connect to the wires physically. 11:59.320 --> 12:03.320 You get that kind of stuff when you run the car. 12:03.320 --> 12:05.320 Or when you do something in the car. 12:05.320 --> 12:08.320 It can be that you pull the handbrake, and then you see something changing. 12:08.320 --> 12:10.320 Or that, you see something changing. 12:10.320 --> 12:18.320 And so you need to start reversing all these messages, because what you see there is you see the idea of the message. 12:18.320 --> 12:25.320 You see the frequency at which they're emitted, and then you see all the bytes that they basically contain. 12:25.320 --> 12:30.320 And you have to figure out which bytes are basically linked to what behavior. 12:30.320 --> 12:33.320 If you light the beam, something will happen. 12:33.320 --> 12:35.320 If you open a door, something will happen. 12:35.320 --> 12:37.320 And you have to figure out, figure this out. 12:37.320 --> 12:48.320 And funny enough, when you start doing that kind of work and that kind of cumbersome, but fun work, you start also realizing that there's very little conventions in the end. 12:48.320 --> 12:56.320 An open door on a VW car would be represented by a byte of, I don't know, FF, for instance. 12:56.320 --> 13:00.320 But it's going to be 43 for another component. 13:00.320 --> 13:05.320 So on and off, we'll have completely different values based on the component you look at. 13:05.320 --> 13:10.320 So standardization is not really what we can see at least from cars of that era. 13:10.320 --> 13:17.320 This is not that great, and the impression we have is that it's still not that great today, but we don't really know. 13:17.320 --> 13:21.320 We will see what we try. 13:21.320 --> 13:32.320 So yeah, doing this, you basically start figuring out what the car is saying, or what a specific component is saying. 13:32.320 --> 13:37.320 And then you have to put these information into some kind of files. 13:37.320 --> 13:46.320 The standard in the industry is DBC files, which you can find some requisitories off, because people have been doing this kind of stuff. 13:46.320 --> 13:55.320 For some time now, I'm just mentioning the open DBC repository from Commodore AI, because it's an interesting repository. 13:55.320 --> 14:03.320 And it gives you an idea of what DBC files look like, kind of a nuggly format to be on us, but that's the industry standard. 14:03.320 --> 14:14.320 And we decided to make our lives easier, and we built our own library to deal with all of that in a lecture in the language that we use for our firmware. 14:14.320 --> 14:30.320 And what we basically did was go for Yamul instead, instead of a cryptic format that DBC is, and what it allows us to do is to specify which frames are being received, which frames are being emitted, which frequency, which IDs and all of that. 14:30.320 --> 14:34.320 And then we can write a mixer code to deal with these messages. 14:34.320 --> 14:42.320 And so that's a complete library that we built specifically for this project, and that you can find on GitHub as of today. 14:42.320 --> 14:45.320 Let's look at what the whole system looks like now. 14:45.320 --> 14:54.320 And on top of it, you see all the components that are from the Nissan Leaf, so we have the actual motor, the inverted charter. 14:54.320 --> 15:03.320 On the left, you have kind of standalone components that we are using connected to a completely different can network. 15:03.320 --> 15:09.320 So you have the battery management system, steering pump and the brake booster. 15:09.320 --> 15:15.320 And on the far right in green, you have all the existing components from the Polo. 15:15.320 --> 15:29.320 So this is the ABS, the dashboard itself, the cluster, the ignition lock, the doors, all of that is actually being connected through specific networks that we created within the car. 15:29.320 --> 15:33.320 And then in the middle, it's everything we built ourselves. 15:33.320 --> 15:46.320 As you can see, so we were in the business of building banking aggregators, now we basically built a canvas aggregator, because you have the vehicle management system on top, which connects all of the different networks together. 15:46.320 --> 15:53.320 And deal with the translation between all of these components and canvases. 15:53.320 --> 16:02.320 And that's really what allows us to have this new brains for the car, and allows us also to use components from different brands inside the same vehicle. 16:02.320 --> 16:16.320 So the vehicle management system is a central piece of the platform that we built, then we have controllers that are running the same code, they didn't use to, but now they are, then the infotainment and all that. 16:16.320 --> 16:20.320 The first step we had to do was to get the leaf motor to spin. 16:20.320 --> 16:23.320 So when you open it up, that looks like that. 16:23.320 --> 16:29.320 So it's not really your usual PC, whatever, it's a bit more difficult to understand. 16:30.320 --> 16:36.320 But we needed to find the correct canvas messages we needed to send in order to make it spin. 16:36.320 --> 16:45.320 So you get the motor, it's pretty heavy, you get it from the delivery man or woman, I don't remember who delivered it. 16:45.320 --> 16:48.320 It's quite heavy, but you don't know if it works. 16:48.320 --> 16:51.320 Until you make it spin, basically. 16:51.320 --> 16:58.320 So fortunately for us, there were a lot of reverse engineering of these motors available online. 16:58.320 --> 17:03.320 And we managed to find the right-hand messages to send in order for it to spin. 17:03.320 --> 17:09.320 And then the next step was to use the gas pedal of the car. 17:09.320 --> 17:14.320 So we could have bought a completely new gas pedal, whatever, but we thought, hey, we already have a gas pedal. 17:14.320 --> 17:21.320 And it's basically just a potential meter, so we can just put an Arduino connected to an Arduino, read the values, 17:21.320 --> 17:24.320 and that will basically be the throttle that we expect. 17:24.320 --> 17:31.320 We expect 100% throttle when, you know, pedal to the metal and 0% throttle when it's completely released. 17:31.320 --> 17:36.320 And so that's basically what we did in first instance, the first iteration of it. 17:36.320 --> 17:42.320 We wired it on a plank and put the wires on an Arduino, we started looking at the values, 17:42.320 --> 17:49.320 and then we created the can messages that were going to be sent to the vehicle measurement system. 17:49.320 --> 17:51.320 And that's how we got started. 17:51.320 --> 17:56.320 You can also see that we have, yeah, I'm not going to try to use the point on this, 17:56.320 --> 18:01.320 but we also have something extra, a debt stage of the project, 18:01.320 --> 18:05.320 which is the gear selector, also on that controller, PR and D. 18:05.320 --> 18:09.320 Because at some point we need to decide in which direction the motor is going to spin. 18:09.320 --> 18:14.320 So we initially added it to that controller. 18:14.320 --> 18:21.320 We also had to build specific controllers for the relays, you know, the contactors, 18:21.320 --> 18:26.320 because for such a motor to actually spin and receive current, 18:26.320 --> 18:29.320 you need to activate relays in a certain order. 18:29.320 --> 18:34.320 So you need to activate some phases in certain orders in order for it to actually spin. 18:34.320 --> 18:40.320 And in order to control that order and to make sure we were not going to fry, 18:40.320 --> 18:43.320 we build those controllers to do this for us. 18:43.320 --> 18:47.320 And in the beginning we used the standard really hat from Arduino, 18:47.320 --> 18:50.320 but somehow they were conflicting. 18:50.320 --> 18:57.320 And then at some point we had the first epiphany, which was the engine spinning 18:57.320 --> 19:00.320 when we were pressing the gas pedal. 19:00.320 --> 19:06.320 And we call it the playing prototype, because everything was literally screwed on a wooden plank. 19:06.320 --> 19:14.320 And yeah, that was the very first time that we got to celebrate because we had something that was actually working. 19:14.320 --> 19:20.320 I should probably mention that none of this would be also possible for us without 3D printing. 19:20.320 --> 19:23.320 So there's been a lot of 3D printing involved in this project. 19:23.320 --> 19:27.320 You'll see it when we move to the more professional prototype. 19:28.320 --> 19:34.320 There's been countless hours also spent on making plans and printing and all of that. 19:34.320 --> 19:40.320 So yeah, kudos to these great printers that we've been using for this project. 19:40.320 --> 19:46.320 So we had to go with something a bit more professional and integrate on the playing prototype. 19:46.320 --> 19:49.320 And that's the version we have right now. 19:49.320 --> 19:56.320 All 3D printed cases, completely new plugs that we had to wire ourselves. 19:56.320 --> 20:03.320 You basically have the VMS on the left, which is Raspberry Pi 4, several canvas hats. 20:03.320 --> 20:08.320 You have also the January controller in the middle. 20:08.320 --> 20:13.320 And somehow that we built because we needed at some point to have multiple ways to connect devices. 20:13.320 --> 20:21.320 And we wanted it to be possible for us to connect new devices really easily to the whole platform. 20:21.320 --> 20:23.320 So we built those small can have. 20:24.320 --> 20:28.320 So the VMS is structured like this, we have build group running on top of that. 20:28.320 --> 20:30.320 We have nerves running year like VM. 20:30.320 --> 20:34.320 Then we have our back end core and API in a mixer in Phoenix. 20:34.320 --> 20:38.320 And on top of that, we also have a dynamic web front end in VUGS. 20:38.320 --> 20:41.320 Why am I mentioning that this is dynamic? 20:41.320 --> 20:50.320 Because the way we created our open vehicle control system is through what we call composers. 20:50.320 --> 20:56.320 And you can basically compose a vehicle from any available component you have in your repository. 20:56.320 --> 21:02.320 So in this case, for instance, we can see that on the left. 21:02.320 --> 21:05.320 This is the composition of the vehicle we have today. 21:05.320 --> 21:08.320 And you can see you have the Polo dashboard. 21:08.320 --> 21:12.320 It has a contact source because they need to know when we flip the key. 21:12.320 --> 21:16.320 The contact source is actually going to be the ignition lock of the Polo. 21:16.320 --> 21:21.320 And then the rotation per minute source, which is a parameter of that specific component, 21:21.320 --> 21:24.320 is provided by something that has nothing to do with the Polo. 21:24.320 --> 21:28.320 It's a leaf inverter that we get the messages from. 21:28.320 --> 21:35.320 So you can really mix and match all types of components in the same vehicle by doing this. 21:35.320 --> 21:38.320 And you just, of course, is still so more to do. 21:38.320 --> 21:44.320 But just have to work the code to write the code that fits to that specific component somewhere in the repository. 21:44.320 --> 21:48.320 And we will go and fetch it. 21:48.320 --> 21:51.320 For instance, in the inverter, the inverter also needs something specific. 21:51.320 --> 21:54.320 It needs the ignition lock, the front controller is a Noviceous component. 21:54.320 --> 21:57.320 So this is from the Noviceous one, of course. 21:57.320 --> 22:01.320 So it allows you really to compose a vehicle anywhere you want. 22:01.320 --> 22:05.320 And we wanted to push it further because we also wanted the dashboard itself, 22:05.320 --> 22:09.320 what we show the interface that we show to also be composable. 22:09.320 --> 22:12.320 So you can have blocks on the page. 22:12.320 --> 22:15.320 And you can show graphs, you can show tables of data. 22:15.320 --> 22:19.320 And you can go get it from any components that is running on your car. 22:19.320 --> 22:22.320 And that basically looks like this. 22:22.320 --> 22:25.320 So. 22:25.320 --> 22:28.320 Hang on, let me go back one. 22:28.320 --> 22:29.320 Okay. 22:29.320 --> 22:34.320 So this this composer dashboard composer interface that I showed you actually 22:34.320 --> 22:40.320 allows you to see all of that data without writing a single line of code. 22:40.320 --> 22:44.320 And you can just hook up to that Raspberry Pi with your laptop or whatever. 22:44.320 --> 22:48.320 The car is actually connected on the internet where we worked. 22:48.320 --> 22:51.320 And then we could just diagnose the car like this. 22:51.320 --> 22:54.320 It was kind of easy to do. 22:54.320 --> 22:55.320 Generic controller. 22:55.320 --> 23:01.320 So in the beginning, as I mentioned, we had specific code running on each of the controllers. 23:01.320 --> 23:05.320 So those controllers are used to gather data from the gas pedal, 23:05.320 --> 23:08.320 for instance, irractivate relays. 23:08.320 --> 23:11.320 But each time we wanted to change something on those controllers, 23:11.320 --> 23:15.320 we had to find the way to update the firmware on them. 23:15.320 --> 23:22.320 So we decided to opt for a different strategy where it's actually the VMS that's going to adopt controllers 23:22.320 --> 23:27.320 and give them the configuration that they are expected to have. 23:27.320 --> 23:32.320 And so all of that happens in a specific adoption process that we created, 23:32.320 --> 23:38.320 allows us to have exactly the same code on every controller to adopt them once. 23:38.320 --> 23:42.320 And then they have the exact purpose that we want them to have. 23:42.320 --> 23:45.320 Now, since this is a pretty old car, 23:45.320 --> 23:49.320 it's not like you have a pretty cool infotainment system on it. 23:49.320 --> 23:55.320 So we thought it would be nice to actually build our own from scratch using nerves as well. 23:55.320 --> 24:02.320 And also we needed a way to select the actual gear, so PR and D in our case, 24:02.320 --> 24:06.320 because we didn't really want to have extra buttons in the car to do that. 24:06.320 --> 24:12.320 And yeah, that was really the idea for the infotainment system at first. 24:12.320 --> 24:17.320 And for this one, we actually went for Flutter to build the interface. 24:17.320 --> 24:24.320 And this is also where we will show later on information about different system statuses in the car. 24:25.320 --> 24:30.320 In the beginning, we had a complete VUGS front end running in a browser. 24:30.320 --> 24:36.320 So it was basically starting in Kiosk mode with cog and webkit, 24:36.320 --> 24:38.320 and it was showing a web page. 24:38.320 --> 24:44.320 And we had poor performance on real time graphs, so we decided to try something else. 24:44.320 --> 24:49.320 And so we went for a specific fork of Flutter by that worked some build route, 24:49.320 --> 24:52.320 which happened to work pretty well. 24:52.320 --> 24:54.320 And we haven't looked back since. 24:54.320 --> 24:57.320 Now Flutter might not be the perfect technology for this, 24:57.320 --> 25:01.320 but at least for what we've been doing on this case is pretty neat. 25:01.320 --> 25:04.320 What does it look like in the car? 25:04.320 --> 25:09.320 It looks like this a bit messy, but on the bottom left you have the VMS, 25:09.320 --> 25:15.320 you have relay boxes, you have the back controller in the middle, 25:15.320 --> 25:21.320 and you can already see the RC car that we've put there in the trunk. 25:21.320 --> 25:24.320 Because when we were testing the RC driving, 25:24.320 --> 25:28.320 the only component we had available was on the car, 25:28.320 --> 25:32.320 so we hooked up directly on the small car. 25:32.320 --> 25:36.320 So the small car was controlling the bigger car, which was kind of fun, 25:36.320 --> 25:37.320 but anyway. 25:37.320 --> 25:40.320 I'll show you a video of this afterwards. 25:40.320 --> 25:43.320 All right, small adaptations we had to do to the car, 25:43.320 --> 25:45.320 really quickly on this one. 25:45.320 --> 25:48.320 I don't know how much you want to know about mechanical stuff, 25:48.320 --> 25:54.320 but in a car of that age, the braking assistance is actually provided by the thermal engine. 25:54.320 --> 25:57.320 In the sense that there's depression that's being caused, 25:57.320 --> 26:00.320 and that depression provides you braking assistance. 26:00.320 --> 26:02.320 But guess what, we don't have a thermal engine anymore, 26:02.320 --> 26:05.320 so you have to find another way to provide braking assistance. 26:05.320 --> 26:08.320 And so it happens that Tesla is a pretty cool component, 26:08.320 --> 26:10.320 which they call the I booster, 26:10.320 --> 26:13.320 just provided 12 volts and it provides you braking assistance. 26:13.320 --> 26:16.320 So just write that and that's it. 26:16.320 --> 26:17.320 Hydroling pump. 26:17.320 --> 26:22.320 Remember that small box on the left in the small architecture diagram I showed you? 26:22.320 --> 26:27.320 This is the Hydroling pump for the steering wheel of the polo. 26:27.320 --> 26:35.320 And it so happens that this kind of device only starts when the engine is running. 26:35.320 --> 26:40.320 Meaning when the diesel engine is at idle RPM level. 26:40.320 --> 26:46.320 But right now we have a motor, an electric motor that has zero RPM when it's powered on. 26:46.320 --> 26:51.320 So we need to find a way to force it to actually start. 26:51.320 --> 26:55.320 And so we have to control it completely independently from what it used to be. 26:55.320 --> 27:01.320 So we had to reverse engineer the canmessages and send the canmessages at the right moment. 27:01.320 --> 27:04.320 Anyway, another thing is building the batteries. 27:04.320 --> 27:07.320 Yes, we didn't buy existing batteries. 27:07.320 --> 27:11.320 We took a use cells and wired it all together. 27:11.320 --> 27:13.320 A pretty dangerous work. 27:13.320 --> 27:18.320 Don't necessarily do this at work, but it took us a lot of time to make this right. 27:18.320 --> 27:23.320 A lot of 3D printed pieces like I said, this plenty of stuff there. 27:23.320 --> 27:26.320 So we have an electric car. 27:26.320 --> 27:30.320 Why don't we transform it into an RC? 27:30.320 --> 27:33.320 We already know we can accelerate, right? 27:33.320 --> 27:36.320 We can send canmessages to make the motor spin. 27:36.320 --> 27:38.320 That we know. 27:38.320 --> 27:39.320 How do we break? 27:39.320 --> 27:40.320 How do we steer? 27:40.320 --> 27:42.320 That's still a question for us. 27:42.320 --> 27:45.320 Breaking actually turned out to be easier than we thought. 27:45.320 --> 27:50.320 Because the Tesla I booster is basically what you have when you have auto pilot. 27:50.320 --> 27:59.320 So it also has canmessages that can control the rate of the fluid that goes through the pump. 27:59.320 --> 28:02.320 And so you can actually control that with canmessages. 28:02.320 --> 28:06.320 And we managed to reverse that as well and figure out what to send in order to do that. 28:06.320 --> 28:11.320 And at some point the breaking pedal started moving based on canmessages. 28:11.320 --> 28:17.320 So that was a second win for us and we were quite happy to see it move. 28:17.320 --> 28:19.320 So breaking that's done. 28:19.320 --> 28:22.320 The next step was to figure out how to steer. 28:23.320 --> 28:29.320 At first we tried to reverse engineer an existing steering column from a more recent polo. 28:29.320 --> 28:31.320 But we kind of failed doing that. 28:31.320 --> 28:33.320 So we decided to just trip away. 28:33.320 --> 28:39.320 It's ICU and serve a motor and just, you know, just put one of ours on it. 28:39.320 --> 28:48.320 So it means that the steering column is now controlled by an Arduino with a controller like a generic controller, like I explained before. 28:48.320 --> 28:53.320 And this allowed us full flexibility over how to actually control it. 28:53.320 --> 28:58.320 So we don't depend on the manufacturer's ICU anymore for this component. 28:58.320 --> 29:03.320 So it's kind of different from the approach we took for the break booster or from the motor. 29:03.320 --> 29:08.320 But at some point we needed to find a way to move forward with the project. 29:08.320 --> 29:15.320 And clearly at that point, reverse engineering the steering column was not working in order advantage. 29:15.320 --> 29:20.320 But the card that you saw driving before in the beginning of the talk has this. 29:20.320 --> 29:24.320 So if you have the remote control, flip a switch and then you can control it. 29:24.320 --> 29:29.320 Unfortunately we didn't have a Dolorion at hand for those that remember back to the future. 29:29.320 --> 29:33.320 I would have been way cooler to do this with a Dolorion. 29:33.320 --> 29:34.320 But yeah. 29:34.320 --> 29:43.320 Because it's pretty much the same experience that Doc is actually having when he drives it in the 2.0 mic. 29:44.320 --> 29:45.320 What now? 29:45.320 --> 29:46.320 Okay. 29:46.320 --> 29:47.320 So we're at this stage. 29:47.320 --> 29:48.320 The car is there. 29:48.320 --> 29:49.320 It's stored. 29:49.320 --> 29:53.320 We're going to test it somewhere in the next few months of the facility. 29:53.320 --> 29:55.320 We need to reserve the place and all that. 29:55.320 --> 29:57.320 Because we can't drive it in an open road. 29:57.320 --> 30:02.320 So we will drive it in a parking lot that's dedicated to testing vehicles. 30:02.320 --> 30:07.320 And also there's a track next to it that we can use as well at some point. 30:07.320 --> 30:08.320 What are we doing now? 30:08.320 --> 30:10.320 What are we still working on? 30:10.320 --> 30:19.320 The OVCS mini turns out that doing remote control stuff and autonomous driving on a wanted car is not necessarily super safe. 30:19.320 --> 30:24.320 So you probably have to test things on a smaller vehicle that will not kill you. 30:24.320 --> 30:33.320 And so everything you saw all of the stack, the hardware and software stack has been put into an RC car that's much, much smaller. 30:33.320 --> 30:36.320 But it's actually the same stack. 30:36.320 --> 30:41.320 And it's just that the composer for the vehicle is slightly different because there are some components that don't exist. 30:41.320 --> 30:43.320 There's no inverter and all that. 30:43.320 --> 30:46.320 But the platform allows you that flexibility. 30:46.320 --> 30:49.320 It also uses can bus to work. 30:49.320 --> 30:56.320 We are also working on a Ross 2 bridge for the car so that you can tell it or you can hook it up to any. 30:56.320 --> 30:59.320 Any stack that exists for for Ross 2. 30:59.320 --> 31:03.320 The cool thing about this bridge is that it's not going to run Ubuntu. 31:03.320 --> 31:10.320 There's an actual a lixir client for Ross 2, which runs on build route. 31:10.320 --> 31:12.320 So it's much simpler. 31:12.320 --> 31:14.320 We have something that's super efficient. 31:14.320 --> 31:18.320 And also working on a on a perception stack at the moment. 31:18.320 --> 31:21.320 We have a few things running right now. 31:21.320 --> 31:26.320 But again, all of these things at some point need to be put together and make sense. 31:26.320 --> 31:28.320 So right now we have. 31:28.320 --> 31:30.320 We detect objects. 31:30.320 --> 31:34.320 We send the objects to the bounding boxes in the classes to Ross 2 topic somewhere. 31:34.320 --> 31:40.320 So all of this is in progress and will at some point be put together into the same bridge. 31:40.320 --> 31:44.320 There's already rudimentary lane detection as well. 31:44.320 --> 31:46.320 That's it's not shown here. 31:46.320 --> 31:52.320 Another component because why not is an OVCS remote because instead of using a 31:53.320 --> 31:56.320 RF or a transmitter or a muffling remote. 31:56.320 --> 31:58.320 We would be nice to have a multi protocol. 31:58.320 --> 32:02.320 Something we can put data back into onto a screen. 32:02.320 --> 32:06.320 And it's also just cool to figure out how these things works as well. 32:06.320 --> 32:09.320 So again, all of this built with open source software. 32:09.320 --> 32:10.320 I didn't show much code. 32:10.320 --> 32:13.320 Sorry about that for those who wanted to see it. 32:13.320 --> 32:16.320 But everything is all the code is open source. 32:16.320 --> 32:19.320 For those asking, yeah, but what about the hardware? 32:19.320 --> 32:20.320 There will be as well. 32:20.320 --> 32:22.320 But you know that's documenting hardware is more. 32:22.320 --> 32:25.320 Takes more time than documenting software. 32:25.320 --> 32:28.320 Even if software is not well-documented at that moment. 32:28.320 --> 32:34.320 But we will definitely release all of this very soon in the next few months. 32:34.320 --> 32:37.320 So that everybody can see what we did, how we did it. 32:37.320 --> 32:42.320 And have an idea of how fun it is and what fun they can have. 32:42.320 --> 32:46.320 And again, small reminder, none of this is road certified. 32:46.320 --> 32:53.320 It might be one day, but you need to look at this as a massive hobby project that we've been working on. 32:53.320 --> 32:57.320 So if you want to find out more about this project than follow the progress. 32:57.320 --> 33:02.320 You can look it up on GitHub right now or just drop us an email. 33:02.320 --> 33:09.320 So thanks for attending this talk even though this part has been recorded afterwards. 33:09.320 --> 33:12.320 And thanks for your attention. 33:16.320 --> 33:45.320 And so on the RC car, somehow the picks started going crazy and it was basically spinning the motor out. 33:45.320 --> 33:51.320 So the RC car just went completely to the wall. 33:51.320 --> 33:55.320 If that would have happened on the big car, we would have been in trouble. 33:55.320 --> 34:00.320 So that's exactly why we won that small scale car. 34:00.320 --> 34:02.320 There's a question over there. 34:02.320 --> 34:05.320 I think you really want to ask it. 34:05.320 --> 34:08.320 Oh yeah, there's questions everywhere. 34:08.320 --> 34:09.320 Sorry. 34:09.320 --> 34:10.320 Okay. 34:10.320 --> 34:11.320 Two microphones. 34:11.320 --> 34:13.320 I only want speaker. 34:13.320 --> 34:17.320 I struggle to find time to play with an ESP32. 34:17.320 --> 34:20.320 Can you tell something about the human resource part? 34:20.320 --> 34:28.320 How much time of your life you put into this thing and how much budget you actually put into this project? 34:28.320 --> 34:31.320 So we didn't work full time last year. 34:31.320 --> 34:35.320 Again, we were supposed to also take a rest and all that. 34:35.320 --> 34:42.320 But we did work about three to four days a week for the whole year, three people. 34:42.320 --> 34:46.320 So I wouldn't say it were long days, by the way. 34:46.320 --> 34:49.320 It was not like a 9 to 19 or something like that. 34:49.320 --> 34:53.320 But you know, much order and sometimes with a beer. 34:53.320 --> 34:59.320 So yeah, I would say maybe one, one and a half full time over a year would be pretty. 34:59.320 --> 35:03.320 Well, one and a half full time over a year would be a pretty fair statement. 35:03.320 --> 35:09.320 In terms of money invested, there's about 20, 25k that we put in the car. 35:09.320 --> 35:13.320 Nobody would be crazy enough to put that kind of money on a 07 car. 35:13.320 --> 35:17.320 But that's pretty much where we are right now. 35:17.320 --> 35:24.320 Everything self-funded and that's why we also are going to take some contracting work this year. 35:24.320 --> 35:27.320 Because at some point we need to eat and pay rent. 35:27.320 --> 35:29.320 But yeah, all right. 35:29.320 --> 35:35.320 I have a question regarding something like this that was showed. 35:35.320 --> 35:39.320 Let's say that there is an out-to-performed. 35:39.320 --> 35:47.320 So how do you make sure that this kind of out-to-doesn't trigger something that would be reverse engineering 35:47.320 --> 35:51.320 for even something different, right? 35:51.320 --> 35:58.320 I mean, I think it opens a dialogue here because some stuff like those parts and everything 35:58.320 --> 36:02.320 needs to be connected out-to-check for whatever reason. 36:02.320 --> 36:07.320 And how to make sure that the process that's involved in there, 36:07.320 --> 36:10.320 I mean, we don't deserve existing things. 36:10.320 --> 36:14.320 That's the one to bring or trigger some kind of weird, 36:14.320 --> 36:19.320 another issue or maybe, you know, reverse engineering even, 36:19.320 --> 36:22.320 and fucking sell out to Diedon China. 36:22.320 --> 36:24.320 Yeah, you don't. 36:24.320 --> 36:26.320 It can happen. 36:27.320 --> 36:32.320 Fortunately, it's not like we have the only issue we have stripped from a component 36:32.320 --> 36:34.320 is on the steering column we used. 36:34.320 --> 36:39.320 The system was bound to both the vehicle management system and the Nissan leaf can, 36:39.320 --> 36:42.320 is that the problem with the diagram is on its own can. 36:42.320 --> 36:44.320 It's on its own can network. 36:44.320 --> 36:48.320 Yeah, sorry, the arrow is a bit weird, but it is completely separate. 36:48.320 --> 36:50.320 It's not linked to any other. 36:50.320 --> 36:55.320 And the reason we have all these can networks is because manufacturers may have chosen the same 36:55.320 --> 36:58.320 ID for messages, but to do different things. 36:58.320 --> 37:01.320 So you don't want them to clash together on the same network. 37:01.320 --> 37:04.320 So you have to keep them separate. 37:04.320 --> 37:06.320 All right. 37:10.320 --> 37:12.320 Yeah, okay, cool. 37:12.320 --> 37:15.320 I was wondering, you mentioned, like, safety is the first priority. 37:15.320 --> 37:17.320 And you're going to do, you know, testing. 37:17.320 --> 37:21.320 What fill save do you have in place, like, for when you do this test, 37:21.320 --> 37:27.320 that if something goes wrong, you have a means of stopping the car or, you know. 37:27.320 --> 37:36.320 So you don't see it that much on the last picture, but anyway, it, it crashed. 37:36.320 --> 37:41.320 Anyway, there is a big red switch inside the car. 37:41.320 --> 37:48.320 If you trigger it, so if you just trigger it, it shuts down all of our platform. 37:48.320 --> 37:52.320 And you're left with two things, two mechanical things that are still existing. 37:52.320 --> 37:56.320 You're left with the steering wheel, that still works. 37:56.320 --> 37:58.320 And the brakes, that still works. 37:58.320 --> 38:03.320 You don't have any throttle anymore, and you don't have any component of hours that still runs. 38:03.320 --> 38:05.320 It just shuts down everything. 38:05.320 --> 38:07.320 And cuts, of course, the high voltage. 38:07.320 --> 38:10.320 So it really cuts everything that can go wrong. 38:10.320 --> 38:14.320 And you just have to find the way to steer the car and break. 38:15.320 --> 38:18.320 That's just safe as we can be while testing. 38:18.320 --> 38:27.320 And also, you don't notice it, but there's a small thing you can pull at the side of the car. 38:27.320 --> 38:29.320 Sorry? 38:31.320 --> 38:36.320 Meaning that when we, when we will test it for autonomous driving at very low speed, 38:36.320 --> 38:41.320 we will be next to it, and if anything goes wrong, we just pull it off. 38:41.320 --> 38:47.320 But it's likely we're going to have somebody in the car in the first steps for these ones. 38:47.320 --> 38:52.320 I know that's a pie because it's a shape, and you can easily do that. 38:52.320 --> 38:59.320 But what if it's really hot or cold, and you know, by will not act and run in the same way, 38:59.320 --> 39:07.320 when it's X now, when you go up and down and for the car, you need something that can stand like a plus for 40 or in, you know, 39:07.320 --> 39:09.320 or minus something. 39:09.320 --> 39:17.320 So did you think about to use some other board that's also free there to, and that can stand that? 39:17.320 --> 39:19.320 Well, yeah. 39:19.320 --> 39:24.320 We looked into other alternatives, but it felt like the easiest to find was Raspberry Pi's an Arduino. 39:24.320 --> 39:28.320 So we went for this, the cheapest as well for what we were doing. 39:28.320 --> 39:31.320 But yeah, there are other boards that are probably better. 39:31.320 --> 39:34.320 We just haven't done a complete analysis of it. 39:34.320 --> 39:37.320 So yeah, it might break if it's too hot or too cold. 39:37.320 --> 39:38.320 Yeah. 39:38.320 --> 39:41.320 We won't drive it at minus 40. 39:46.320 --> 39:47.320 Hello. 39:47.320 --> 39:49.320 Thanks for the presentation. 39:49.320 --> 39:51.320 I got a question about it. 39:51.320 --> 39:57.320 Again, it's something that's literally a lab on wheels, a more than something that's meant to be driveable, 39:57.320 --> 40:02.320 but a different platform, different car, would have made a different, of course. 40:02.320 --> 40:07.320 But with the car that we had, which is a city size car, this is how we built it. 40:07.320 --> 40:10.320 We could also have put the batteries somewhere else, whatever. 40:10.320 --> 40:14.320 But anyway, there's always, you know, something you could have done differently. 40:14.320 --> 40:19.320 But yeah, a lot of this, Louis also welded all the welding job. 40:19.320 --> 40:27.320 We had to weld for the batteries to be slid into something that is closed, I would say. 40:27.320 --> 40:33.320 We have aluminum boxes that were welded so that we can put the battery back. 40:33.320 --> 40:37.320 It's there are other people who can't cause it a similar. 40:37.320 --> 40:42.320 Well, yeah, first we need to test it, see how fast it drives. 40:42.320 --> 40:47.320 And then, well, first we need to make sure it's safe and then see how fast it drives. 40:47.320 --> 40:52.320 So yeah, that's definitely one way this could go. 40:53.320 --> 40:58.320 We will see, honestly, it depends on what the testing says. 40:58.320 --> 41:01.320 Okay. 41:01.320 --> 41:08.320 Hey, working with Alex here, you often hear letters crash for a handling. 41:08.320 --> 41:10.320 Does it work with Sierra? 41:10.320 --> 41:17.320 No, we don't want the car to crash, so it doesn't give a different meaning to the saying. 41:17.320 --> 41:19.320 No, of course. 41:22.320 --> 41:25.320 Hi, so one more harder aside. 41:25.320 --> 41:27.320 Yes, a harder aside. 41:27.320 --> 41:33.320 Have you, like, you said that maybe someday when road legal have you given it any, like, research, 41:33.320 --> 41:39.320 how much would it take to actually make just a software, you know, road legal. 41:39.320 --> 41:41.320 So we haven't. 41:41.320 --> 41:44.320 Hey, thanks for your presentation. 41:44.320 --> 41:48.320 And I have a bit of a matter of question, which is, you mentioned that, 41:48.320 --> 41:50.320 we could have tried this, we could have tried that. 41:50.320 --> 41:54.320 And that really interests me because this is a very open end and project. 41:54.320 --> 41:59.320 And since you're doing it, like, kind of on your own, on your own schedules, you could have gone in any direction. 41:59.320 --> 42:00.320 Yeah. 42:00.320 --> 42:04.320 So how, but yet you still got a lot done in a very short amount of time. 42:04.320 --> 42:08.320 So how did you kind of keep your eyes on the ball and not get focused on, like, 42:08.320 --> 42:12.320 let's do something else completely different and abandon this project? 42:12.320 --> 42:16.320 I think we were always focused on the goal of making the wheel spin. 42:16.320 --> 42:19.320 So there's a few milestones that we wanted to reach. 42:19.320 --> 42:22.320 First was make the wheel spin. 42:22.320 --> 42:26.320 And so the first step to make the wheel spin was make the motor spin with the pedal. 42:26.320 --> 42:27.320 So something like that. 42:27.320 --> 42:32.320 So at some, we got lost also in many directions all the time. 42:32.320 --> 42:36.320 But we always came back to, okay, what's the next goal? 42:36.320 --> 42:40.320 So when we felt we were going a stray, each in or on direction, 42:40.320 --> 42:43.320 we would sit back and say, okay, but what's the next goal? 42:43.320 --> 42:47.320 And said, okay, we want to, we want to show this at this conference. 42:47.320 --> 42:48.320 We want to show that at this cup. 42:48.320 --> 42:53.320 So also, again, conferences like this or talks are a good way to, you know, 42:53.320 --> 42:58.320 kick you in the right direction, because you want to show something at that specific moment. 42:58.320 --> 43:01.320 Sometimes you make it, sometimes you don't, but at least it gives you a goal. 43:01.320 --> 43:06.320 So you can go astray, but then if you come back to something tangible, 43:06.320 --> 43:12.320 and you have a reason to then it's easier and easier. 43:12.320 --> 43:15.320 Okay, thanks for the presentation. 43:15.320 --> 43:20.320 I was wondering how long it takes until the system reacts to import, 43:20.320 --> 43:24.320 as you're booting Linux and then the virtual machine. 43:24.320 --> 43:26.320 So. 43:26.320 --> 43:30.320 Well, we haven't noticed any significant latency, honestly. 43:30.320 --> 43:34.320 So, but we haven't driven it at 120 kilometers on the highway. 43:34.320 --> 43:40.320 So we don't know yet. 43:40.320 --> 43:47.320 But what we can see is that the frames are being emitted or received at a pretty decent frequency. 43:47.320 --> 43:50.320 So it's not like the footprint is making it completely insane. 43:50.320 --> 43:56.320 So it's not like messages we're expecting every 100 milliseconds or suddenly coming every 200 or 300. 43:56.320 --> 44:01.320 So. 44:01.320 --> 44:04.320 At the boot time, okay. 44:04.320 --> 44:08.320 The boot time that we have with this, so the the infotainment you saw, 44:08.320 --> 44:12.320 we're about six or seven seconds to boot the screen. 44:12.320 --> 44:14.320 The VMS is like is running on the right. 44:14.320 --> 44:17.320 So it's running on the Raspberry Pi 5, the infotainment system. 44:17.320 --> 44:19.320 So it's a bit faster. 44:19.320 --> 44:21.320 And then the VMS is pretty much the same. 44:21.320 --> 44:25.320 I would say 10 seconds to boot everything up. 44:25.320 --> 44:30.320 But the idea is that it would be always on just like on a on a on a on a 44:30.320 --> 44:33.320 V, you know, most of the system they stay always on. 44:33.320 --> 44:36.320 And when you get into the car, then it lights up. 44:36.320 --> 44:40.320 But there's a lot of things that stay up in the car. 44:40.320 --> 44:42.320 And that would be the idea. 44:42.320 --> 44:45.320 Since we have this huge massive battery, we can just, you know, 44:45.320 --> 44:49.320 take a bit of juice out of it to keep our platform always on. 44:49.320 --> 44:54.320 And what turns it off is actually the big red switch itself. 44:54.320 --> 44:58.320 And if you keep it always on, then you don't need to care about boot time. 44:58.320 --> 45:00.320 It's only when you switch it on again. 45:00.320 --> 45:02.320 That's that you have that problem. 45:02.320 --> 45:06.320 But we want to stay below the 10 second boot time for everything. 45:06.320 --> 45:08.320 That's what we aim for. 45:08.320 --> 45:12.320 And for now we've managed to do it. 45:12.320 --> 45:14.320 Hi. 45:14.320 --> 45:15.320 Okay. 45:15.320 --> 45:18.320 What did you do with the entire instrument cluster stack? 45:18.320 --> 45:22.320 As I noticed, it wasn't, it didn't seem to be there anyway. 45:22.320 --> 45:23.320 Oh, yes. 45:23.320 --> 45:24.320 It's still there. 45:24.320 --> 45:27.320 And you can watch the both talk from yesterday. 45:27.320 --> 45:32.320 Because we also wanted to make it as seamless as possible. 45:32.320 --> 45:35.320 The changes that we made to the car. 45:35.320 --> 45:38.320 So that it's actually the RPM needle. 45:38.320 --> 45:41.320 Shows the RPMs of the leaf engine. 45:41.320 --> 45:43.320 So we basically went there. 45:43.320 --> 45:48.320 And the next step when we have everything right with the battery is also to use the fuel gauge 45:48.320 --> 45:53.320 to show the percentage of battery left on the existing fuel tank gauge. 45:53.320 --> 45:57.320 We've already figured out how to use a digital potential meter to do that. 45:57.320 --> 45:58.320 A lot of curiosity. 45:58.320 --> 46:01.320 That's the RPM range bits. 46:01.320 --> 46:05.320 Because it's a little a lot about the electric engine now. 46:05.320 --> 46:10.320 But in most retrofit work, you will see people changing it. 46:10.320 --> 46:14.320 Basically, which we find a bit stupid because it's already there. 46:14.320 --> 46:16.320 And it's kind of interesting. 46:16.320 --> 46:18.320 If you go to that level of integration. 46:18.320 --> 46:19.320 So it's really digging. 46:19.320 --> 46:22.320 You don't have a separate button to start the electric motor. 46:22.320 --> 46:25.320 It's really the polo ignition log. 46:25.320 --> 46:28.320 So you do it like you would do on a diesel engine. 46:28.320 --> 46:31.320 And it starts the motor, which is pretty sweet. 46:31.320 --> 46:35.320 Because then it means you don't have an external system to start everything. 46:35.320 --> 46:37.320 If you have the can messages for all of that, 46:37.320 --> 46:40.320 we can trigger whatever you want on your own platform like this. 46:40.320 --> 46:42.320 And it makes the integration as seamless as possible. 46:42.320 --> 46:45.320 And it's almost like the car was electric to begin with. 46:45.320 --> 46:48.320 Which is really nice when you're inside the car. 46:53.320 --> 46:54.320 Is it? 46:54.320 --> 46:55.320 Yeah. 46:55.320 --> 46:57.320 Thank you very inspiring talk. 46:57.320 --> 47:00.320 Is there any practicality like of your project? 47:00.320 --> 47:02.320 What is the goal? 47:02.320 --> 47:03.320 Because it's open-ended. 47:03.320 --> 47:06.320 You said it's never going to be road safe. 47:06.320 --> 47:09.320 Software is pretty customized. 47:09.320 --> 47:10.320 Never. 47:10.320 --> 47:11.320 Just never say never. 47:11.320 --> 47:13.320 For now it's not. 47:13.320 --> 47:14.320 But what's the goal? 47:14.320 --> 47:16.320 That's quite interesting. 47:16.320 --> 47:17.320 Question. 47:17.320 --> 47:21.320 I wouldn't say we have a definitive answer to this one yet. 47:21.320 --> 47:23.320 We don't know yet. 47:23.320 --> 47:25.320 We know what milestones we want to reach. 47:25.320 --> 47:29.320 What we want eventually is to see a drive alone around the track. 47:29.320 --> 47:31.320 That's what we want. 47:31.320 --> 47:35.320 And we will put the work into making it happen. 47:35.320 --> 47:39.320 What's interesting, however, is everything we've learned along the way. 47:39.320 --> 47:47.320 And give us ideas of things we could be doing either with the platform or completely different things we could be doing in products we could be building. 47:47.320 --> 47:55.320 So the goal there was really learning new stuff, figuring out what we want to do in the next years. 47:55.320 --> 47:58.320 We haven't answered all the questions yet. 47:58.320 --> 48:05.320 But we have accumulated quite a lot of knowledge and interest in different topics that we didn't have before. 48:05.320 --> 48:06.320 So it was really the opportunity. 48:06.320 --> 48:10.320 It's something you never have an opportunity to work for on a production environment. 48:11.320 --> 48:19.320 Even research projects that universities don't necessarily go that far in these in that direction. 48:19.320 --> 48:21.320 We're using existing forests and all of that. 48:21.320 --> 48:24.320 So it's completely, yeah, a bit crazy I would say. 48:24.320 --> 48:27.320 But what's, yeah. 48:29.320 --> 48:31.320 Hey, yeah, amazing talk. 48:31.320 --> 48:32.320 Thanks. 48:32.320 --> 48:35.320 I wanted why you used sales for the battery. 48:35.320 --> 48:38.320 Was all the did you buy them with the motor? 48:38.320 --> 48:39.320 Was the motor used as well? 48:39.320 --> 48:44.320 Or did you buy like a new motor and for some reason have used battery sales? 48:44.320 --> 48:53.320 So one of the things we did and that we would have done differently is that we kind of went to an EV shop online. 48:53.320 --> 48:56.320 We said we want to buy a kit to convert a car. 48:56.320 --> 49:02.320 And it turned out that we didn't really receive a working kit. 49:03.320 --> 49:08.320 So the BMS or the battery management system was not the right dimension. 49:08.320 --> 49:11.320 They gave us also an ECU that was not compatible with. 49:11.320 --> 49:12.320 So it was a mess. 49:12.320 --> 49:15.320 And we fortunately we got some money back there. 49:15.320 --> 49:20.320 And for the battery, we were supposed to get an existing Nissan EV battery. 49:20.320 --> 49:25.320 And it was not coming in and somebody said, okay, you need to find an alternative. 49:25.320 --> 49:29.320 And they sent us a battery sales from a Nissan NV200. 49:29.320 --> 49:31.320 So that's what you see there. 49:31.320 --> 49:35.320 And we're much the same battery sales but configured differently. 49:35.320 --> 49:39.320 And so what we noticed is that we just got the sales. 49:39.320 --> 49:42.320 There's no bus bar, no wires, nothing. 49:42.320 --> 49:46.320 So you have to, you just see this when you start working on it, right? 49:46.320 --> 49:48.320 Say yeah, where are the bus bars? 49:48.320 --> 49:51.320 And so you have to figure out all of this by yourself. 49:51.320 --> 49:56.320 It would have been much easier to have a battery pack already done to cable. 49:56.320 --> 49:57.320 Let's go. 49:57.320 --> 50:00.320 But we didn't get that. 50:00.320 --> 50:07.320 So I wouldn't say we were scam because we got something for a fair price. 50:07.320 --> 50:11.320 But that was definitely not a ready-made kit that you can just put in a car. 50:11.320 --> 50:16.320 And I think that's also something when you start converting cars or if you want to do converting cars, 50:16.320 --> 50:17.320 EV cars is a hobby. 50:17.320 --> 50:20.320 Be very careful with what you can find online and all that. 50:20.320 --> 50:26.320 You will receive stuff but maybe not everything will work together as it's meant to. 50:26.320 --> 50:28.320 And so we had to figure this out. 50:28.320 --> 50:34.320 We still have to figure out quite a few things also with the BMS and all that. 50:34.320 --> 50:39.320 So yeah, it's, it happens. 50:39.320 --> 50:41.320 Hello. 50:41.320 --> 50:43.320 Very, very lovely. 50:43.320 --> 50:44.320 Thank you. 50:44.320 --> 50:50.320 I was also reverse engineering campus on Volvo V50. 50:50.320 --> 50:53.320 And I wonder, like you did much more. 50:53.320 --> 50:57.320 So like I published the IDs on GitHub. 50:57.320 --> 51:04.320 So do you know maybe if the same model of the same year, like do they reuse the IDs? 51:04.320 --> 51:09.320 Does it make sense to publish the IDs you found? 51:09.320 --> 51:17.320 VW has had different platforms, platform meaning electrical and communication platform. 51:18.320 --> 51:21.320 The one from that poll, I think is PQ24. 51:21.320 --> 51:23.320 I don't think it anyway. 51:23.320 --> 51:25.320 So I think they had 2425, 26. 51:25.320 --> 51:30.320 And the PQ platforms were just evolution on the same one. 51:30.320 --> 51:38.320 And then they went for MQB or something like that, which changed completely because they went for a platform that was a five in number correctly. 51:38.320 --> 51:46.320 Maybe I'm saying stupid things but which was supposed to be the same for all of their brands like Scoda and a VW and all that. 51:46.320 --> 51:49.320 That was the MQB platform. 51:49.320 --> 51:55.320 So there is already some documentation available for this platform, but not everything is there by the way. 51:55.320 --> 52:00.320 But what you will find is people that recorded a canvas trace along the way. 52:00.320 --> 52:07.320 So they just drove the car, got a canvas trace, all the messages for, I don't know, five minutes ten minutes and then they dumped it online. 52:07.320 --> 52:09.320 So a lot of that you can find. 52:09.320 --> 52:20.320 By the way, if you if you plan on reversing can on a car you're working on, don't remove the diesel engine first because when we started the project we're like, yeah, let's just remove the engine. 52:20.320 --> 52:27.320 And then okay, now we need to reverse engineer the can but we don't have probably what is the central piece of the car. 52:27.320 --> 52:35.320 So yeah, anyway, so you can find some documentation online, but it's not perfectly structured in a way that you can easily find it. 52:36.320 --> 52:46.320 Sometimes you find information in forum tracks, sometimes you find information on get up, sometimes you find information by just talking to people on a random channel somewhere. 52:46.320 --> 52:51.320 So it's it's a bit of a rabbit hole. 52:51.320 --> 53:04.320 Just like I mean, if if you reverse engineer for example one component of this car, maybe you know like from your like community feedback that the same component. 53:04.320 --> 53:16.320 And in the other car does it use the same IDs or it might be a completely different, do you need to read their engineer like in every car the same component. 53:16.320 --> 53:31.320 I we don't know all components of all brands it might be that sometimes they use a standard configuration for the component we don't know what we have seen however is that the I booster from Tesla so the server break system. 53:31.320 --> 53:44.320 It's also using other cars so it's a Bosch component basically it's a Bosch component that is used in several cars but it seems like the ideas of the messages are different so they probably program them. 53:44.320 --> 53:48.320 Somewhere at the manufacturer or even maybe. 53:48.320 --> 54:05.320 But they program them at Bosch we don't know maybe they send what needs to be on there without the IDs and all that to Bosch and then Bosch can figure it so not always but in some cases it might be the same messages we don't know there's so many components there that we. 54:05.320 --> 54:06.320 Last question. 54:06.320 --> 54:17.320 So are there any elements that you're looking for contributors for and are you looking for contributors to the project at all and if yes. 54:17.320 --> 54:21.320 Or what what could people do to help. 54:21.320 --> 54:26.320 Okay, I was not expecting that question. 54:26.320 --> 54:34.320 Why if you want to get started send an email we would happily help you get started and figure out what's missing or how to. 54:34.320 --> 54:40.320 Right now what we've done is mostly limited to the components we had access to. 54:40.320 --> 54:49.320 If someone could just build you know components from other cars just by connected to the canvas and creating all the things that would be quite amazing. 54:49.320 --> 54:55.320 To add new components to the platform because it would make it much more open. 54:55.320 --> 55:03.320 There's also ongoing work we said the Ross to bridge and all that is still stuff to figure out. 55:03.320 --> 55:10.320 So there's plenty of things to do but it's not like we have a list of things that contributors could jump into. 55:10.320 --> 55:23.320 We would love to keep working on that project as well even do paid work based on that project but for now it's just again we're just showcasing it to see if anybody would be interested in working with us. 55:23.320 --> 55:34.320 Or a company that maybe willing to pay us to keep some of that work going and use it in their own products or use it on their own use cases but so far it's not the case. 55:34.320 --> 55:37.320 So we're taking on some other contracting work this year. 55:37.320 --> 55:43.320 But yeah anything that can help us keep you know keeping this alive and keep working on it would be great. 55:43.320 --> 55:49.320 And if you want to help just send us an email and we will happily see what can be done. 55:49.320 --> 55:53.320 I'm over time so thanks everyone. I'm still around so if you have any more questions. 55:53.320 --> 55:55.320 Many, many, thanks.