 Our speakers we have Mr. Tanat Thong-Wu Thaisi from NASDA and STDA, he's an innovative consultant to NASDA, another government agency focusing on science and technology, and he also focused his domains in blockchain, fintech, and AI, so please welcome Mr. Tanat, thank you very much. Good morning everyone, so my name is Tanat from NASDA, actually I'm no longer with NASDA as of today, so this is our latest development, I'm now with the Royal Thai Army, I'll be working on cyber security, but I'll be helping out the autonomous vehicle community to develop the actual AV solutions in Thailand, and although we're not aiming for the level four or level five which is like fully autonomous solution, but at least hopefully we can develop a level three, there's a technical term called ADAS, Advanced Driver Assistance System in Thailand for the next few years. So let's see, what have we got, well, since I was from NASDA I have to promote a little work that we've done with the private companies, so that's Sancerie, one of the larger property developments in Thailand, so Sancerie Cruiserie Ventures which is their corporate venture capitalist, has signed an MOU with NASDA, so the gentleman, the second from your left, is one of our Executive Vice President, he looked after the automotive session and they decided MOU along with the Serive Avenger CTO, Kunjirapat Chumjir Sak of Serive Ventures. By the end of this year, Sancerie will launch a township called T77 along Sukumvit 77, it's a new concept of having the large community together, so the size of T77 is about 40 acres and the longest road, longest path from the front gate to the deepest side of T77 is about two kilometers long. By the end of this year, Sancerie will remove the headquarter along with 3,000 staff to T77, so they are looking for solutions to help with the transportation mobility of their staff as well as the people in the communities. At T77, there are going to be about eight condominiums varying in sizes, they are primary school, they are community mall called Habito, they are working together with Habba to have that community mall. So they allow us to use that property to be a sandbox and test the AV solutions. In the long term, go that NASDA along with our new ministry, which used to be known as Ministry of Science and Technology, but now we are known as, in short, it's called MESI, so M-H-E-S-R-I, that's Ministry of Higher Education, Science, Research and Innovation. Our three-year goal is to have the autonomous shuttle service, but again, we are not aiming for level four or level five, we are going to have at most level three, partly because of the regulations that we have to work on, Ministry of Transport as well. We have the MOU, the Ministry of Transport, and they are looking up to our prototype before we can proceed with the proper regulations. Now, in terms of talent development, we want to bring something fun to Thailand. Amazon Web Services, during their measure and no event, re-invent, normally happening in November, each year in Las Vegas, they launch this new toy called DeepRacer, essentially it's a building hardware equipped with this latest technology for a toy car. People who want to get involved only need to implement one of the AI techniques called reinforcement learning, so DeepRacer allows the shorter time to learn this RL or reinforcement learning model in a fun way. We are looking to bring DeepRacer League in Thailand. At the moment, there are about 22 physical leagues across the world and there are six virtual leagues. Amazon Web Services, the public sector staff is working with one of the university, which is KMITL, King Mungut Institute of Technology, to develop academic programs on autonomous vehicle. Basically, they can train their students to win a race and then we can use that as a magnet to bring the DeepRacer physical league to Thailand in next year also. This is an unrelated development. I will show you shortly how it is involved with another Thai universities. So, NVIDIA, again, Dr. Phu Xiong is using the cloud computing computational power, DGX1, for the edge computing for automotive. They have this product called Drive Platform and the latest on the most powerful edge computing for automotive is Drive GX Pegasus. The rating is about 320 tops. Tops is different from teraflops. Teraflops is used for floating point operation only, but tops you can use for calculation for both integers and floating point. So, 20 tops, that's equated to be about 80 teraflops. So it's about one-tenth of the processing power of DGX1. But remember that it is an edge device, so it is quite powerful. One of the universities in the southern Thailand, that's Prinsong Khad Yai campus, they've been experimenting computer vision with Autonomous Vehicle, their edge computing right now relies on a smaller-scale edge device from NVIDIA as well called Jetson. Jetson, unfortunately, is designed for CCTV or the traffic controller board. It's not designed for automotive, so it gets overheated very easily. So NVIDIA launched this product instead, and they have partnered with one of the well-known very long-established car makers that's Volvo to have the edge computing installed in the so-called Autonomous Vehicle solutions. So Prinsong Khad University is going to procure that very, very shortly. NASA is one of the contact points that we want to smooth things through between dealing with the vendors and the academic sectors. So hopefully we can see the drive platform being experimented in Thailand soon. I have to note that drive platform is only available in eight countries, not including Thailand, so we have to go through NVIDIA Singapore, prove ourselves of our enthusiasm to work on Autonomous Vehicle, and we have to get authorization from NVIDIA Japan who look after a drive platform in the Asia Pacific before we can procure that. So that's a quick recap. We talked about T-77, DeepRacer, the major partner is going to be King Mongkut in Ladgabang. We are also talking about one of the holding companies of the telco or telecommunication company. They want to sponsor this event. Again, it came through Channel of NASA, so NASA is going to be the intermediary to connect them together. Infineon is one of the hardware manufacturers. Infineon is also having an autonomous vehicle lead of their own. So Infineon is already a partner of KMITL. So what we are thinking is after launching DeepRacer lead, DeepRacer only involved with the software layer which is reinforcement learning, and the hardware itself is pre-built by Amazon Web Services. But in the long run, if we can get the students to work on the hardware itself by using Infineon technologies, we can combine both reinforcement learning models through Amazon Web Services Cloud or other Cloud provider. And with the hardware, that students can customize on their own using Infineon framework. Okay, we talked about NVIDIA Drive. Open Motors is a startup from Silicon Valley. They graduated from Y Combinator in 2014. The founders worked with Pyrene Farina in Italy. So let me just show you their kind of design. So the outer layer, the look and feel of their vehicles are very modern and unlike many other AV solutions. And that concept is to have a modular design. Think of your phone. When your battery is kind of depleted or break down, instead of changing your entire phone, you can just change the batteries. If your screen is broken, you can just change the screen. So concept of Open Motors is everything is modular. You can adapt everything according to your needs. For example, this vehicle is designed for a passenger's car. But if you want to use that for logistics, you can change the top. You don't need any seats for passengers. You can change it to the holding space for cargo. So that's the concept of Open Motors. So we are working with them to bring Open Motors concept to Thailand. We are connecting suppliers of different parts from the steering, the braking, the powertrain to the seats, the outer layer, the lighting. So hopefully in the next few years, we might see very nice looking electric and autonomous vehicle on the closed area. I left out the part about public road yet because we don't have the rules and regulations in place yet, but hopefully we will have that soon. Just back shortly, we are also working on the consortium, partnering between private and public sectors, and also the regulators. So we want to have this roadmap about if you want to have AV work running on this closed area or even on the public road in the next 10 or 20 years, what we need to do. So we are working with a few governing bodies, public sectors and universities to work on that white paper, which will be publicized once it is ready. Okay, a few points about open source, open data. For example, the positioning of AV, we rely on GNSS, GNSS is a more generic term of GPS. GPS is a US name, really, it's not a generic term. Generic term is GNSS or Geo navigational satellite system, it allows you to find the positioning of each object, in this case, this autonomous vehicle. The more accurate solution of GNSS is RTK or return kinematics. It allows you to pinpoint down to millimeters accuracy. So open source library called RTK Lib, so that's part of the open source initiative. The second part, federated learning, this is a new concept of if you have the private data installed in your own devices and you do not want to expose that data, how you can train and improve your AI or deep learning model. Federated learning allows each device who volunteer to participate to train or update the model in their own device and then they send only the weighting, the adjusted weight back to the central server and the central server will federate all these models together. Actually from memory tomorrow, being the second of October in the evening, I think there's a meetup. The first meetup on federated learning, just go to meetup.com, look at federated learning. The organizer actually work with Andrew Trask of Google, Andrew Trask is one of the pioneers in federated learning with Google DeepMind. You can learn more about how federated learning work, AV is one of the use case for federated learning. We will have a new role, new jobs role called data labeler. So when you have this raw data, you need to tell, you need to label them so that you can train the AI model accordingly. This can be a new jobs for the industry. Every Amazon Web Services, they have this concept that you can employ a physical worker, a real human worker to do your dirty work, those tedious mundane work called mechanical turds. So you pay this very low fee and you get actual human to do certain things for you. But it's very simple task. For example, labeling data is one of that task. In China, you have these chop chops, right? People just go in, sit in front of their computers, label the data all day long and you use this label data to train your AI model. So autonomous vehicle, we need to rely on this label data as well. And we have open source data set. This is a very good direction for the autonomous vehicle community. By far, there are three major players, FTIF. You might not know FTIF, but you might have heard of Neutonomy. Neutonomy is autonomous vehicle startup or more towards robot taxi. So autonomous taxi service in Singapore and they have been acquired by Aptif. They open data set called new scenes. Also algoverse by algo.ai, they just opened up that recently. And also Waymo. If you are interested in training images for autonomous vehicle, you can look up this treated data set which is available as an open data. Interesting development. Let me see. Okay. I don't have much time. Now, organizer Misery introduced me to Mapillary. It's the open data street map. You can contribute, you know, if you love traveling and you carry your smartphone along, you know, take photos, upload it. You can help the community build another open data for street maps as well. Okay. Open interface software alliance, OSA. It's another open source initiative for 5G connectivity. Even the connectivity side, some of the libraries are proprietary, but the major player knows that open source community will drive the adoption of this technology. So they allow certain part of the codes to be open source. Okay. Let me skip to Sony. Sony is building an autonomous vehicle with a new concept, with entertainment in mind. So people sitting, the passengers sitting inside can see the interior to be the entertainment while they are moving along. And the outside, you can also have another business model. You can display advertisement as well for the other people. So this is going to be a new business model for the new era once we have the autonomous vehicle running on the street. Effectiva is a company specialized in computer vision. In this case, it can be applied to detect the emotion or the state of the chauffeur or the drivers, whether or not they are sleepy, whether or not they are using their mobile without concentrating on the road. This is especially harmful if you are not on the level five fully autonomous vehicle. It still requires driver attention on the road. Perceptive automata claims that it can understand human behaviors. For example, humans surrounding you, either the pedestrian or the drivers in the car nearby you. Perceptive automata claims that they can anticipate their decision, what they are going to do next, which allow autonomous vehicle to act, your own autonomous vehicle, the one you are sitting in, to act appropriately and improve the safety. And we have standards, normal automotive standards, functional safety. ISO 262, that's for the system to fail. Can it fail safely without having caused any accidents? Whereas the another ISO standard, 21448, it check whether or not the intended functionality actually is safe. Because even though it works as planned, but the design itself might not be safe. So these two standards are a compliment. And autonomous vehicle, the most important components, because we want to focus on safety. So the braking system is very crucial. And at the moment, there are two manufacturers that have this extra safe braking system, so Bosch and Continental. And the Ministry of, you know, MESSI, Higher Education, Science, Research and Innovation is working with one of them. Actually, we want to work with two of them, with both of them, but only one of them got in touch with us. So we are going to trial one of their products on our autonomous vehicle prototype. And we have to go through other standards like Euro NCAP and see how Singapore, the Land Transport Authority, LTA, they have this technical reference, TR-68, which actually allow them to skyrocket in the autonomous vehicle readiness index to be number two in the world according to KPMG. They're only second to the Dutch or the Netherlands because of this TR-68. So let's go through quickly. So there are four parts in TR-68 about basic behavior of standalone autonomous vehicle, not connected. Connectivity will come in the later part. So essentially how it interpret the road signs, the lanes, the marking signals and so on and how the policy and rules are designed in the framework. Safety, Elon Musk, Lex Fridman of MIT, people who work in the autonomous vehicle industry in the US. People agree that the most important thing about autonomous vehicle is to promote road safety. So safety is a very crucial part for anyone who want to get into this arena. So in Singapore, they focus on this part and they focus on how AV can be operated safely on the public road. In Thailand, I foresee the soonest that we can see that happen in probably 10 years. In Singapore, they might see it in three years. In this part, they also have various use cases for Singapore. So if Thailand is to follow suit, we're going to have to define our own use cases and ensure that there are road safety suitable for those use cases as well. And there are many system level safety, functional and operational. And if you see the third bullet about people involved with the autonomous vehicle, whether or not they are vehicle developer, SI or system integrator or even the operator, system operator. In Singapore, they say they need to be competent organization, they need to have a proper quality management system and also it needs to be supported by competent personnel. So this is something that our government needs to work on as well, how to certify the competent organization and what is the proper quality management system and how to certify competent personnel to support it. Cybersecurity is another crucial example. Even in the western world, people in automotive do not know much about cybersecurity but you know that in DEFCON or Black Hat Con, people hack Tesla, people hack Volkswagen all the time. But these hackers, they don't come to automotive exhibition or conference, they only gather themselves in this kind of hackers related computer geek conference. So we need to connect both worlds together and ensure that autonomous vehicle manufacturers are also concerned about cybersecurity standards and best practice. The parts like, there are parts with the cyber physical meaning that there are part that are traditional automotive and combinational part that interact together and that computational part can control the physical part of the automotive. That needs to be tested against any cybersecurity threat otherwise, you know, if you've seen the fast and the furious, I'm not sure which, maybe part seven or something, I think the bad guy hack into, they use the zero day threat to hack into autonomous vehicle and get them to work according to their malicious plan. But if there are proper controls and checks in place, that can never happen. Okay, let's go through this quickly. This last part on TR68 is about connectivity. So again, in Thailand, it's going to be a long, long time because we need infrastructure. For example, Dr. Puchong mentioned about 5G testbed with the Cassette State University in Siddhartha campus. They are the very first site that have this 5G infrastructure but it's being rolled out across the country very slowly. Hopefully, there will be more testbed for 5G which we can test on the 5G application. We might see the latest new that one of the minister has control and a vehicle from far away. The vehicle itself is in Hat Yai operated by Prince Om Kha University. The minister, Dr. Sobit, he was in Bangkok and he controlled his vehicle during the heavy rain situation. So apparently they have rain attenuation on the signal as well but they managed to have this successful launch, successful test by the 5G network, also sponsored by AIS. Again, this is for the connectivity level 4 and level 5. It's going to be a long, long time. Thailand also needs to work on the data infrastructure as well for that to happen. Just a quick, this is my last full slide, AVRI, Autonomous Vehicle Readiness Index according to KPMG. So they, how they indexed the top 25 capable countries on AV, they break down into the four pillars. For the first one, apparently it's policy and legislation. Rest assured, the Ministry of Transport, there are subdepartments like Department of Land Transport, there's Departments of Highway, Departments of Rural Roads, and Office of the Transportation Policy, OTP, they are working on the Autonomous Vehicle Policy. They are involved with our round table discussion on AV. But we need to see the actual prototype before they can draft the legislation and rules accordingly. Technology and innovation, apparently there's a buzzword with the public organization called PPP, Public-Private Partnership. So we are promoting that in Thailand as well. So we need to partner with, you know, under private companies to work on the AV solutions. And we need more investment on the AV-related firm. And our researchers, our scholarship students who came back from overseas are bringing their knowledge from their professors from the U.S., Europe, Japan together to build this localized solution for AV. Also, we also need the more market share on AV as well. Because even though we don't need electric vehicle, but electric vehicle will actually help with development of AV because autonomous vehicles rely on sensors. Autonomous vehicles use electricity, so these sensors can be installed very easily on AV. But you can work on these sensors or the autonomous function on the traditional internal combustion engine or ICE, or those cars that use petrol. It can be done as well. Infrastructure, this is not only the data or IT infrastructure, not only telecommunication infrastructure, but also the quality of roads. Rest assured that, you know, if you live in Thailand long enough, you know the quality of our road. But rest assured that the relevant departments, they are working to improve the quality of our roads so that AV can be adopted sooner rather than later. Also, logistic performance index. Logistic industry will be one of the demand side or the demand pool that see the AV in trucks or in cargo transportation to be adopted in the near future. I think this is my last slide. Public awareness. We see in, I think, Pittsburgh, way more operated on public road on Pittsburgh. But certain people who might be in technophobia, they actually grab baseball bat and smash on the car because they do not understand how autonomous vehicle can help improve their quality of life. So again, we need to engage the users, the general public from the very early stage through various measures in this slide. Again, I do not have my contact details, but if you like, you can reach me via line application. My ID is social. So that's me. S-O-C-I-A-L. That's my line ID. I'll look up my full name in LinkedIn. I'm more than happy to connect with you. And if you are interested in the roundtable talk, you do not need any technical background, even being an observer, we are more than welcome because, again, this last slide is about engaging the general public to be aware about the upcoming autonomous vehicle solutions. So thank you very much. Thank you so much, Mr. Tanat.