Keynote from Dr. Huei Peng at 20/20 CCAT Global Symposium
So, John Quinn this morning gave a very good keynote, speech and he talked a lot about the current status of kinetic because on my talk will. Be a very, brief run connecting, our focus, on the, automotive. Part especially, what. We are considered, to be most, important. Research. In this areas and I'll spend. Some time talking about future track. So. How I. Minimize. This screen. So you're not, ok, so this is a current trend everybody no case try to connect your commentary. Electrified, an MCD. Only. Focused, on connect. Automatically. Gonna care about steering, electric, device that also. Very competitive. And people, are already doing it you can buy or, you can use service, in. Sharing Electrify, so we focus on the kinetic, automated, especially, the pre competitive. Landscape. In. The. Goal of course is always the. Society, is very important, we always, want to improve the safety, energy. Congestion. Accessibility. Of the society, by. Studying. And use of the, emerging, technologies, such as the fe phi g AI, h, computer. So. This is a very quick, example. Of what we do at M city we have been trying to do education. Understand. What is the relationship. Between. Connecticut. O mated and how do they help. Automated, driving this is a good example we, actually demonstrated. This to, you. Know many visitors, lastly. We have fewer we have about 450. Groups, of people visiting, M city the, year before it's about 600 groups, this. Is one of the demos we have been using, to show that, connectivity. Doesn't, need large site while. Humanized, camera, Dalai that they only lie. Outside so, in certain driving. Situations. Connectivity. Is very important, and also, in. General the relationship. Between kinetic, and automated, vehicle is that. You. The automatic, vehicle will be generating, a lot of data much. Of the data Michael, to the cloud to help us understand, how people, drive their vehicles, and how the transportation. System work, in. The meantime. Connectivity. Was helping, to have an online of sight sensing. Information. Can. Provide map, and traffic information to, the automatic, vehicles, and, can also be the enabling. Messer, for over-the-air, update, not here the connectivity, is in more general, including, possibly, to be, so. DSRC. CB 2 x also, 4G. 5g LTE, for, example. What. What is the most. Important, characteristic. Of AV is that they, really address the top three human behavior, risk factors, including. If, we look at the crash, data, alcohol. Use speeding, distracted. Driving are the top three factors. Contributing. To crashes. And very costly, there. Are other factors C values and non use and motorcycle. Helmet non use are also, there but those are not addressed, by AV necessarily. But maybe is definitely, helping so, we all know distracting help distracted. Driving is very important, safety is critical. Also. A V promise, to help. With aging, society we, all know that you. Know younger countries. Like African. Countries, India a very, young but many other countries like Japan Italy, are. Very old the. Population. Distribution, is more like a kite shape instead. Of a pyramid shape and. Automatic. Vehicles certainly help whether, it's a partially. Automated or, autonomous, driving post are very, very harmful for aging Society I. Think. I may be a preaching to the choir here by, basically. Saying we, always, argue. That we are starting automated. Vehicles not autonomous, vehicles. And automatic. Is really a more, evolutionary. Approach. So. We are now at level zero everything is done by human. We're moving to level four level five and I agree. With Steve balsa who is a panelist, this morning who said that level. Five is. Aspiration. Or asymptote. Today's. Vehicles are all only up to level four and chances. Are they will stay as quote, unquote level, four for a very very very long time but. So it's. Really all, the way from zero to four, instead. Of a jump all the way to four, and. Automatic, vehicles can have. Three kind of functions, guiding function call Tyler or chauffeur function, the. Consumer, is that the general public always, I excited, about the, chauffeur, function, but, reality is that. These. Technologies will, help to improve safety in edgy, society, by moving, up the ladder so to speak by. First providing, Guardian, functions through say a ep autonomous, emergency braking, and. Then level one level two as, a co-pilot, the, human, driver is still taking the responsibility. But. Eventually will. Be moving to a higher level like say level, four. There. Are plenty of, you. Know vehicles. With these. Automated. Features, you, can buy vehicles with a B or, adaptive. Cruise control lane keeping, assist. To. Test out autopilot, GM, supercruise, and so forth all, the way to highly. Automated. Vehicles. You. Have probably, seen the Navia chateau, being deployed, at. European campus, from June 2018. To, December. 2019. Or, you, have heard them say mayo clinic is now using the same Navia shuttle. Transporting. Coffee knighting, tests, imposed in medical.
Supplies. Now. The deployment. Or test, drive up. In, California, Phoenix, and so forth have been suspended. By. Way more uber. Probably. Also heard that the zukes lay off many. Of their test drivers, so. Obviously, Kovan, I think making some. Impact with the development, of this EKOS but. Level 4 is with us even though had. Limited. Field, geofence, limited. Speed, weather. Limited, and so forth and. Are. They effective, there, are plenty of studies that. Demonstrating. That a be, on. Commercial. Deployment has, people been tracking, the, performance of a be demonstrating. That they are very effective and. Many other aid assistance, were. Also studied and they people show that yes statistically. There, are very strong. Evidence. That they are helping, to improve safety so all is good. Sometimes. When, you are designing, something. You. Know we're, working with a human we, need to worry about how, do you use. The information to. Collaborate. With a human, for, example parking, sensor, only means that they only beep. The. Human driver the. Real camera means that you only display. The camera image and, it's, up to the human driver to react now, the arrow. Shows. That the, study performance. Is, in the range sometimes people, human driver become over. Trusting, or they misuse, or they don't understand. How, the system work and sometimes. The number of crashes increased, whether they decrease, already. We're on the very, right, hand side if you see that. If. You are not. Only having in the sensor. Ultrasonic. Sensor assurance radar camera, also. You have automated braking. Statistically. We don't see any increase. Of number. Of crashes the. Best. The. Best case is you, can reduce the crash by about 80%. Now. I don't know how the video, quality plays, out here. Now. Level 2 vehicles are relatively, new in, many places in the world are, developing. Test. Standards. Especially. Out the EuroNCAP has. Developed. On concept for testing. In this case, your. ACC lane-keeping. Systems, and they, are testing cutting caudal. Situation. For example in. In. The u.s.. IHS. Is also, doing. Something right, now everything, is a little bit ad-hoc, there is no cinder process, there's certainly no. Nothing. In the work that will make this into. A FMVSS. Federal, Motor Vehicle Safety standard turn, that you, have heard many times today during, the symposium. There's, also no. Near-term. Prospect. Of having this becoming. Part. Of the entire, process, yet, but. A B may become either. Mandatory. Or there, will be a standard, testing, procedure, pretty, soon ACC. Line keeping will still take a little bit a little while. So. This is an example of, people. At IHS. Testing. This, is in this case a lane keeping system if. You pay attention to the video on the lower left corner you. Can see that the vehicle sometimes did, a very good job tracking, the lane line and was, able to work, even during this Lane split line merge situation, and, this is this video you can see that. Sometimes. They come, encroaching. On the lane lines but. Overall, this. Is how people, have been testing, some, of ad hoc there's notes in the procedure. In. Nitza, is. Liberty, RTC. They. Have been also be testing, for pedestrian. Cutting. Other kind, of customers. But. I know if we summarize on, the. Current. Situation there is no. Federal. Or even. Any. Other level, standard, test procedure for.
Level 1 level 2 because, people. Are still trying. To understand, how best to test, those results, and report, the results in. A way that is easy to understand, by the general, public. We. All know that this, is an example of. Test. A fatal, crash test. Autopilot, is a level 2 vehicle, this, is the California, fatal crash where the Apple engineer. Was killed. NTSB, just published, a. Report. On, this particular crash, there. Are like 14, NTSB. Active. Cases for. Tesla recalls, and. This is one they just published the results, at the end they. Cited, many factors, a driver destruction of course is still a major factor and then, insufficient. Fit, oversight, and, they, talked about event. Data recording. So I, agree, with one comment, from this morning's, panel, that there. May be some. Urgency, in terms of the. Funding. The regulation. Defining. The regulation, need for this energy technology. How. About level for driverless vehicles, this is the Whooper. Vehicle, that killed a pedestrian, who, walk bicycle. Similar. Things people, are trying to imagine how they are tested. But there's no standard, testing. Procedure. Many. Of you probably already. Studied. The 2019. California. DMV everything. Occasional reports. Interesting. This year, we, see many more, companies. That are approaching, the. Line of. 110, thousand miles between, this engagement. And. Then there are still many other vehicles. That are still below the 100 mile, or disengagement, number, but. This is not the perfect index, however they are this. Is probably. One. Of the most prevalent, yardstick. We have it's not perfect, place it's, there. Now. In, terms of level vehicle. Testing M City has proposed, a concept, we, say that, basically. We need to test many, miles but, we need to be smart we don't, want to do put force, miles. Want, to do accelerate, evaluation. We, also propose a set of behavior. Competence. Scenarios. We, also argue. That depending. On your deployment. Your operation, designed domain you, need to define corner, cases that, truly, reflect, the, worst-possible-case, toward. The corner of your operational, design domains, so, I don't have time to talk about the detail, of the ABC, guys but, that's not the point anyway for today's talk, now, right. Now of course we, all know that the, sensors everybody, was basically, using the same set of sensors, camera. Dolly, the GPS, HD map some. Are starting to use other type, for example thermal. Camera some. Are now using. Different. Pattern of light are scanning but, there, are many challenges. Cause. Automotive. Great reliability. How. Do you do sensor fusion, how. Do you be, especially, using, artificial, neural network, how do you, integrate. The pros and comes the strengths and weakness of these sensors, so, that the perception. And. The, detection, are very robust that's still the key challenge, there. Are many companies. Or, you, know University, researchers. Demonstrate. That they can do pretty well on, a. Nice clean, day demonstrated. By the middle-left. Photo. But, the, reality. Is that there. Are many other challenges and, having the perception, system working. Robustly. In all the other pictures you see on the slide is is challenging, we. Also know. That the radar struggle, in certain, situation, lida, is also, struggling other situation. So, since. They're feuding is important, and right, now most of the perception.
Study. Was, basically. Dedicated to, the camera. The. New network has demonstrated to, do, very well in. Comparison, to the traditional rule. Base for pattern recognition. Education. Based, sale. And detection will be equal detection, however. In. Comparison. Less. Was done unlike, us in the radars, that are. Still. Demonstrating, that they give you a robust, measurement, but the, rest loschen was, not quite there. So. I'm just under all this situation we all we have seen many of the past. Activity, is showing, that we, have many more to be, done and what, does m-city, do today. So. M, City is a public-private, partnership we, focus on the positive. Research, and deployment work so, earlier, this morning there. Was a question. Asking. Why pon we focused, on level-1 level-2 and, our, answer, to that and, the questions always we, need to think. About the future and we need to work on pre competitive. Area. So level, 1 level 2 is pretty competitive, and. Car. Companies are already selling, those devices, or vehicles. So. Maybe, the testing, the evaluation. Part is pre competitive, but. That's quite, relevant to what we do in, terms of the level for testing anyway. So. M city does a laboratory. Development. Many of you have seen the M city test facility, we, also argue, that having. The living lab in an upper in, Southeast, Michigan and, going forward we are going to Detroit, to deploy. There, were four shout, out to try to improve, the mobility, situation. In in, Detroit, we. Also support, many research, projects, we also. Support. Education, outreach activities. Many. Of you have seen in City so, I hope this is nada. I don't. Want need to spend too much time here, we. Have 18 acre test facility, but most importantly, we, develop. Augmented. Reality test. Capabilities. Working. With the professor's ahem you do and also. In, terms of simulations. That we work with many simulation, companies, right. Now seven. Commercial. Simulation. Software. Package, plus Scala and sumo they, all have em city test facility. As, a virtual, environment and, we think that by, combining, the stimulation, and experiment. We can develop the, much faster, and it, also happened, to be the first test facility, having 5g. Capabilities. We, first have Verizon, 5g. More. Than 1/2, years ago now we have the second, generation the exactly, the same Phi, G devices. That they deploy. In many. Cities to them anyplace in the world, so. These are the names, of the seven software, companies, that have, modeled, EM City all. The four pictures you see here are not, photos, they are all simulation. Results. And, the. We. Work with stand so that we argue. We need to have, both real, testing, and that digital twing available. To our company. Members. This. Is the video showing the example. Of augmented. Reality testing. In, simulation. Happening, in EM City on the, left hand side that's a simulation, done by leasing, on. The right hand side on the top you can see a camera. View from inside the cabin and, also outward looking. As. Well as on the lower right corner you, can see the view from the vehicle, in, the vehicle we only see a black green.
Black Yellow. Black so forth in, the reason you see much nicer rendering. Of the outside. World we. Also have we are the only. City. Or deployment. Project. In the world that, basically, collect data and, share, data with our members. We. Start deploying, this, in June 2018, and, these, are examples of the videos we captured, and the, example, of how the data can be used, understanding. The possession. Level. For shuttle interactions, how, people, choose, to use. The shackle, the, number on the right hand side are, basically. Showing. That the sitting, preference, of. Passengers. When. They come, to use our shadows, we have collected, more than 16,000, trips. And, 2300. Hours of video and again, despite. Of. There are dozens, of, deployment, all around the world we, are the only place, that. Is. Collecting. A shame that they talk with our company, members. Overall, we have collected, a lot of data, MCT, hesitate a garage, we, are now featuring. Fifty. Fifty. Seven sets, of data many. Of them are not, unique. Like. For example the you, can see that we label, the lane lines the, because. Traffic. Signs, and. So forth not. Unique however, we are work. Very hard and we understand, that collecting. Data and labeling data sharing data is very. Can. Be very pre-competitive. Right, the, new network can be done by the companies, but we are the place that make, data, collection. Sharing. Widely. Available to, many of our company, members, this. Is one example of, key database. That's collected by country. Professor. To say King. Ser and. Debby, buzina working. On the safety pilot model deployment, project. And, over. The years we have collected up. To about 2800. Vehicles about, more. Than 50, Minnie Mouse I know my number is always. Lower. Than the real number but. More than 50 million miles of driving and we shared. The PSM and some of, the state data with our members, and they can be used for many purposes. And. Let me spend, the last few minutes talking. About the future, trend, of. Italian. Nobility. Connected. Vehicle, to cloud is, LTE. 4G so, is. Prevalent, and. Many. Cases in the future will be gone through the smartphones this is something that we don't need to do this is competitive, already, happening. V2x. Is one area that we can work together. Unfortunately. Because of the FCC, NPRM, there, will be significant. Uncertainty, even for. The next two to three years again, John quant who from, Ford gave, a very good review, of the current status so if you don't have a chance to listen to his keynote. Speech in, the morning, I strongly, recommend that you watch the. Video that's, captured. By sleek ad. Now. In terms of automotive vehicles, a b and other ideas. ADA's features will become very very popular, because, we, don't really need to offer only we can, use some help by, the guardian or compiler. Functions, even. Data recording, compromise can, provide more insightful, information and, we mean that it is time to really rethink, what. The future, event, data recording. Messages. Or information, should be and, how we. Use that in a way that, is a wing, wing for the OEMs, and for the. Drivers. Level-1. Level-2 vehicles. Will grow very quickly because. The, sensor, are becoming so. Much more reliable and. Cheaper, and, we'll. See a lot more vehicles. With this features, label. For vehicles were cross still grow slowly, in niche applications. For. Example the city center's afraid, and goods delivery, first/last, malleability. And, other. Utilities. Such as snow plowing or garbage. Collection, and so forth but they will remain a in, the niche application. In limited, geo-fenced, area. For a while. In. Terms of data this is really the area that we need to think and work together the. Reason is that data. Is the, core. Of making. The city smart, and people's, life better. There's, the, biggest challenge, is always the balance, between privacy, and, public good I think. We can do that just. People need to somehow. Not taking, that absolute. View even, a Governor. Cuomo, is now talking about using. Big data to do better job in. Terms of tracking he. Was talking about more. Testing, and also tracking. Career did a very good job in tracking, Taiwan. Singapore are. Doing, a much better job in tracking, the coronavirus. Patients. And. They. Are able to control the situation so, with. That understanding I hope we, also strike a better balance between, privacy and, probably. Good. And. Companies. Are using data for preventative, maintenance and.
People. Are now saying imagining. Usage, based tax users, based insurance and. So forth. People. Are reading take another look at event, data recording. What. Is the new. EDR. For CVS. Futures. Cafe, standards, also people, are imagining that. We don't use a drive. Cycle, based test, results, but instead there, are a lot of data right, there already. People. Are also saying hey you. Know one, of the major things about driving is about for detection. The. Way forward and I think this particular, plot, from a TSP, report from the Tesla crash, is. Is, the indication, of the future so Tesla capture, a lot of data they, actually know exactly right, before the crash what. Happened and even in this particular case they, can say hey you you, the Apple, engineer. Who draws through the same place last. Time you pay attention and. The. System function. Exactly the same way but, you were able to hand, out take over but, this time you didn't because, you were playing video games they can even have, enough data to say things, like that now, I'm not arguing who's right who's wrong, I'm just, arguing that going. Forward we really, need to rethink how. What. Kind of data we collect, all such data how. Do we use them for the, better society, oh good. So. How long before we have a vaccine for the coffee 19 there are a lot of things. That, we really need to understand, including. Basia vmb on the, attention, of the federal government is on the virus economy, may be related, making. Even though there are news for different bills. People are thinking about, getting. Junk when is a very good job talking, about that this. Morning but, I think the that. It will not be the focus of attention, every. Testing, deployment. Will also slow and you can see that you. Know uber, Ramos, looks lay off their people, and so forth and although there will be fewer, trips miles by the general, public which is a problem, for. The industry. We. See a lot of hesitancy. To use public transportation, some. Places were reporting, job. By, 90, percent of. Using. Public transportation. Including. Right hell I never, lived in. A short time the. Situation. Is, that we, definitely will see an impact of. Incarceration. I. Reported, 80 percent, drop in fabric, 2020. And about. 40-some. Percent for. The first quarter, in, China we. Think that there will be similar, impact. To the car sales in other places, too in, the medium term because people, hesitate. To use the public transportation, there. May be actually, a bounce-back of, more. Individual, car sales but. I also output that as are certain, and. Going. Forward there will be in pretty for people, goods across borders. International. Of managed take and that is a problem for mobility, but. What what are the opportunities I'll, just point out to two. Things that I think are good. Free. Goods delivery. And other essential, service, I don't, know how many of you are ordering from, Amazon, or Costco, or Kroger, absolutely. I do. Although, many more and it's. A potential, for people to really understand, what is opt, and not saying that oh when. Can you believe we will find the, point is we're. Not shooting. For level 5 level 4 vehicles are equally useful. All. Right so that's my talk I'll, handle, the. Podium. Back to Henry. Or fancy.