In this episode…
Our society is going through a powerful shift as we move from gas-guzzling vehicles to electron-guzzlers. Both established automotive manufacturers and suppliers, as well as start-ups, are developing technologies at break-neck speeds to create the vehicles that will become the standards for the coming age. Along the way, patents will be filed and challenged, and technologies will be founded, stolen, and borrowed.
Professor Mehrdad (Mark) Ehsani and Professor Juan Pimentel have front-row seats to the ongoing show and recently discussed what they’ve observed so far, and what is coming up, as the international free-for-all continues.
Note: Transcript has been lightly edited for clarity.
Dan Rubin: A lot of news coverage seems to assume that electric vehicles are going to achieve a predominant share of the market as they continue to evolve. What are the technical realities, in your opinion, of e-vehicle adoption?
Juan Pimentel: So there are many things that are upsides and advantages about this technology. In terms of the upsides, I would say it’s good for the environment, and also it involves less components, so the system is really less complex if you compare it to a traditional engine vehicle. However, there are also some limitations. There are a number of them. I’d like to just mention a couple. One of them would be the batteries…batteries in terms of the range; in other words, how far they can go on a single charge. And also there are some safety issues associated with the batteries that need to be totally handled. I’m not saying they are unsafe, but there are certain safety issues that you have to be extremely careful the way you deploy these batteries, because they could be a cause for accidents and things like that. In addition to that, you have what is called the charging infrastructure, which is very complex. It is not totally deployed yet. We have made a lot of progress in terms of that. But, in my opinion, this infrastructure will go slowly in terms of the deployment across the entire country or across the entire world, if you will. So that’s my take on that.
Mark Ehsani: An electric vehicle is structurally or architecturally very simple. It consists of a battery and an electric motor and a controller that interacts with the driver and the electric motor. As such, it’s a simple, rugged, robust engineering system that is well understood, and all of the components are actually well-developed. The battery is the only thing that can possibly move the needle in the coming years. But even that has a fundamental theoretical limit. I published a paper on this about 24 years ago about the fundamental limitations of electrochemical batteries. If you build an ideal electrochemical battery from the most electronegative element of the periodic table and the most electropositive element of the periodic table and the ideal electrolyte, and no packaging, the energy density of such a battery is still an order of magnitude smaller than a conventional hydrocarbon gas tank or fuel tank. So, we can make progress, but there’s only so far we can go. So, having said that, don’t expect any razzle-dazzle, revolutionary “we didn’t know where this one came from” thing. It’s just not in physics. So, where are the breakthroughs? The breakthroughs are in fulfilling the potential of the electric vehicle as a product; that is, better product design and development.
There are several advantages to electric vehicles, as Juan mentioned. And there are two fundamental limitations that cannot be overcome. One of them is range: for a given weight of energy storage device–gas tank versus electrochemical battery–an electric vehicle will never meet the range of a conventional vehicle. I fill up my gas tank in my Jeep Grand Cherokee once, and I can drive 700 miles. If we do that with an electric vehicle, the mass of energy storage is going to be many, many times bigger than gas. That’s the first fundamental limitation. The second one is, fundamentally at least, with conventional batteries–the type of batteries we consider for vehicles–the discharge time is tens of times faster than the charge time. So, in a conventional vehicle, I put 24 gallons of gas in the tank. That takes about five minutes. I charge the tank in about three megawatts. A megawatt is the power that a skyscraper needs electrically. So, in five minutes, I charge the battery, and then I spend the next ten hours discharging it. If I take that same ten hours of discharging the electric vehicle on the road, it’ll take longer than that to recharge that battery.
So, what does that say? It’s a new vehicle. And so, it will never fully displace the conventional vehicle. But it can fulfill its potential by better product development. So, this sounds too esoteric. Let me end by saying what does better product development mean? A product that is not a luxury, showoff, feel-good item, but actually is better than a conventional vehicle product. That would be for markets that require four-passenger family transportation for much less cost than a conventional vehicle, say two to five thousand dollars (India, China, Africa, Southeast Asia, South America, and so on–the developing countries); [there], an electric vehicle can be built for that kind of money. So that’s one potential multi-billion population market. The other one is short-range applications. In Taiwan, in Hong Kong, in Panama, in Israel, in many places in Europe, and so on and so forth, and in big cities in this country: for those kinds of markets, this product has to be designed and targeted that way. And let me add that building a $120,000 luxury electric vehicle is not that product. Automotive people are beginning to get it. I read this morning in The Times that Ford is developing, for example, their pickup trucks in the U.S. for this market. Those are the right things to do. Of course, you know, there’s a paradox. There is a pickup truck, a city car, or a country car. With a country car, you’ve got an issue of range. But the right product development is the issue. And frankly, the infrastructure of electric vehicles are not that complex and can be quickly erected. In fact, they require a much shorter supply chain than conventional vehicles. That’s why Tesla could go from zero to full production in a few years. Imagine if General Electric had to be built up from the ground up. It would take a decade or more, and so on.
Dan Rubin: Well, I wanted to stick with you, Dr. Ehsani. You mentioned infrastructure. Related to that, are e-vehicles truly “zero emission?” What are their sources of power, as it relates to infrastructure?
Mark Ehsani: Let me kind of peg that to the current powerplant electrical system, which is fundamentally a hydrocarbon fuel electric power generation. There are some perturbations, some inroads with renewable power sources. For the time being, I’ll leave that alone because we want to talk in concrete steps and then we can speculate about the future. Based on hydrocarbon-based power plants that power our utility grid, I’ve done a calculation. Very optimistically, the fuel-to-wheel economy or efficiency of an electric vehicle is about 28%. So, not only is it hydrocarbon-emitting and all the other emissions from natural gas, coal or whatever, 28% percent can easily be achieved by advancement in conventional vehicles, like proper hybridization and so on and so forth. A conventional, run-of-the-mill, off-the-showroom vehicle today is about 18 to 20% percent efficient. So we have only eight percent increment of emissions; that’s a greenhouse emissions thing there. Now, look, I’ve been doing research on electric vehicles for 35 years, not pro or con. I mean, it’s not a political talk. It’s not a market talk. It’s a technical talk. The limitation that I told you about [with] the electric vehicle is not based on the vehicle. The vehicle is highly optimized. You know, the round-trip energy efficiency of the battery is well over 90%. The efficiency of electric motors is well over 90%. The efficiency of power electronics that controls is approaching 95% or better. Multiply all those things together and get an overall efficiency of 80% or something. But where is the energy coming from? Now, if you say that I’m going to invent a new country where we start with electric vehicles and all the energy comes from renewable energy, then I have to redo that calculation because the only emissions is the cost of the manufacturing of the renewable power plants. Now, having said that, let me add one more thing. Again, the problem is complex. All I can do in a short time is to confuse you rather than solve the problem. The advantage of electric vehicles is that it aggregates the source of energy for transportation. It aggregates it. We don’t have gasoline and diesel for cars, natural gas for heating and illumination of the houses. It aggregates. So, in a country like India, if electric vehicles proliferate because they’re properly designed for that population, by aggregation a country that doesn’t have natural fuel resources can then diversify its sources of electric power and get some of it from hydrocarbon, some of it from wind, from solar, and so on. So that aggregation allows you more efficient transition to renewables in the future. So there are some advantages there. That’s a complex problem. Just because it’s theoretically conceivable doesn’t mean that the legacy power plants are going to lay down and die voluntarily. Nor does it mean that the economies and the technology necessarily are favorable. But it’s conceivable.
Dan Rubin: Talking about the power sources, if we were to speculate that e-vehicles will completely take off, would our infrastructure be able to handle such demand? [SLIDE]
Mark Ehsani: Not immediately, but fortunately this is not going to happen in a step function. It’s going to happen gradually. There’s no fundamental problem or limitation that I can perceive both as a utility power systems engineer or expert, as well as electric vehicle and load type of expert. Those things can track each other fairly well. I mean, one plant may take five years or more depending on what technology you use, nuclear or otherwise, to come online. But when it comes online, it brings one gigawatt online, which actually has a noticeable percentage increment in the capacity of production. And there is one advantage actually: the charging stations, if you will–the gas stations for electric vehicles–are more prolific and less conspicuous. They are in every house, every shop, every parking spot. So, there are some advantages to that, whereas the conventional vehicles have their own supply chain and delivery and charging stations and so on and so forth.
Dan Rubin: Obviously, it’s hard to talk about e-vehicles without talking about autonomous driving. Where does autonomous driving technology stand today?
Juan Pimentel: It turns out that the SAE–which used to be Society of Automotive Engineers, but today is just SAE International–they have developed a model to really classify the automation on levels, and they go from Level 1 to Level 5. So think of Level 1 as what it used to be maybe 20 years ago, where there is no automation whatsoever. And currently today we’re between Levels 2 and 3. And Levels 4 and 5 are the highest levels of automation, and we’re still not there yet. Some people say that it will take us maybe between five and 10 years before we get to Level 5, even though a company like Tesla has been saying that this year they could have achieved Level 5. So a lot of disagreements, but there is no question that achieving Level 5 is extremely, extremely difficult. So today we’re at around Levels 2 and 3.
Dan Rubin: OK. Are there certain problems with autonomous driving that stand out to you?
Juan Pimentel: Yeah, many problems. In fact, the industry, just in the last year, they have realized that this problem of autonomous driving is much more difficult than everybody thought about. Now, to me, that was not surprising because I knew that’s going to be very, very difficult. So, companies, governments–the automotive industry is divided into mainly three players–the OEMs, the Tier 1s that directly work with the OEMs, and the Tier 2s that work for the Tier 1s–so, these institutions or organizations at these three different levels, they play different roles, and they are making big progress, obviously. However, I would say that the biggest inroads are that we’re getting into Level 4 now–between 3 and 4–today. And one of the main issues, by the way, about that is what is called the “perception system,” which is pretty much related to us human beings. I mean, we have our senses, right? Particularly, sight…you know, seeing…very important for cars, because they have to be able to actually detect objects, detect roads. And it’s a big problem because it’s not just one sensor; they have to consider many sensors. And if you think about it, as human beings, it took millions and millions of years really for nature to form our brains, if you will. And somehow thinking that in only a few years you could form the brain of a computer system to actually navigate through a very complex environment, that’s very simplistic. We’re far away from reaching a level of complexity, of thinking of the human mind, if you will, even though there are many, many models, computing models, in the automotive industry, that are similar to the human brain. One of them is called the neural network. The neural network is actually modelled exactly like the brain. I shouldn’t say exactly, sorry, because we really don’t understand how our brain works, but it’s somewhat like the human brain, alright, and there are still a lot of problems with these neural networks, many problems. So, definitely, the perception systems for the autonomous vehicles are one of the big issues that we have today.
Dan Rubin: Could you just elaborate a little bit on the levels that you mentioned? You mentioned Levels 3, 4, 5…
Juan Pimentel: Levels 1, 2, 3, 4 and 5, right? So Level 1, again, is when we had the situation about 20 years ago with no automation whatsoever. Think of Level 1 similar to a cruise control, where there is some automation, but it is very limited. Level 2 would be what’s called an “intelligent cruise control,” where you have a car that is, for instance, going into a lane and still doing some braking for you, so that it doesn’t run into the car in front of you. So that would be Level 2. Level 3 would be “partial automation,” only at some times. But the driver needs to be monitoring the environment at all times and be ready to take over. And that’s a very significant issue. In fact, I expect a lot of, actually, patents and also even lawsuits about Level 3, to be honest with you. It’s a big issue. The reason is that human drivers are involved half of the time or maybe more than that. Level 4 and 5 are similar, but in Level 4 you have what’s called “total automation,” but in a limited domain. That’s what differentiates between Level 4 and 5. In other words, the domain could be: you have full automation, but you’re only in a highway. As soon as the car leaves the highway, then the human takes over. But as long as you are on a highway, you could talk about total automation. And Tesla seems to believe they have Level 4 already. In Level 5, you could actually remove the driving wheel and the brakes from the car. You don’t need it. So we’re still many years away from Level 5.
Dan Rubin: So that was going to be my next question, when do you anticipate that some of these solutions will be available, but you said many years from now?
Juan Pimentel: Level 5, yeah. In my opinion, five to 10 years. We’re still farther away, even though three years ago people thought that this year and next year we will achieve Level 5. There was so much excitement in the industry, it’s unbelievable. They thought that all the problems were basically solved, at least theoretically, and it was just a matter of testing and implementing, but no…the whole industry, we were all wrong. You know, everybody. I’m not talking about just academicians. I’m talking about OEMs, customers, the real players. Even governments. And talking about governments, not the same. Like, the way I see it right now, Europe is a little bit ahead of the rest of the world. They just passed a regulation called WP.29. Very futuristic. And they are going to start implementing it next year. So in Europe–this is for vehicle manufacturers, of course–they have to implement this regulation. This regulation is strictly for cybersecurity, only cybersecurity. And I expect similar regulations for safety also.
Dan Rubin: Do you see the success of autonomous driving as important to the success of e-vehicles as a whole?
Juan Pimentel: Yes and no. It turns out that they are decoupled and independent from one another, but certainly, historically, I don’t know whether it’s great, bad…that’s independent. I think it’s pretty good. What I’m talking about is the development of e-vehicles coincided with the developments of autonomous driving. That’s really amazing. E-vehicles–Mark will probably have something to say about that–are nothing new. In fact, electric vehicles–few people know about this–they came before fuel vehicles. The problem was that we didn’t have the technology to implement them efficiently, so they were on the side, but they were the first ones to be invented, if you will, because it was just such a natural thing to do. Way back, over a hundred years ago, the first vehicles were electric vehicles. But it turned out that the development of electric vehicles had to wait for another hundred years, believe it or not, before they became actually feasible. So this is what I mean that it’s a coincidence because you can have an autonomous vehicle totally on fuel vehicles. Now, having said that, it turns out there are some advantages that these two developments are coinciding because you could probably argue that you could make it more efficient, you can tailor the autonomous driving based on the developments of e-vehicles. That is also true. But this is an accident more so than a planned thing. Having said that, I do see some developments. As Mark mentioned, and I totally agree with that, he mentioned that electric vehicles are far simpler than fuel vehicles. And one could take that fact really to develop autonomous driving in a much better way. And I think Tesla is doing a good job about that, in some aspects.
Mark Ehsani: Number one, as Juan said, the idea of autonomy in a vehicle is virtually totally independent of its propulsion or powerplant. It doesn’t matter if it’s electric, a combustion engine, a hybrid, or anything. So don’t confuse those two just because it’s electricity and computers. A couple of other points: an autonomous vehicle, meaning a vehicle that you can give it or the driving mechanism the destination, and sit back and relax and reliably get there, has existed for over a hundred years. It’s called a taxicab. And so, you have to ask yourself from the user’s point of view, what is the functionality? What is new that it brings to the table? It’s not going to be the fact that it drives itself over there, because it’s been going on in airplanes for years. Most of the time, the role of the pilot in an airplane is monitoring. Even the landing, which is the most complex part of the flight, is now automated, and the pilot is basically monitoring, ready to take over. That tells you a few things. Number one, the human element is always going to be there. So the use of the autonomy is secondary, from the standpoint of safety or fatigue reduction over long trips. And, by the way, I just bought my wife a car, and it drives itself on the highway. It kind of gives you a nasty warning to keep one hand on the steering, but it’s not necessary. It drives itself down the highway just fine. But I have to be there. I don’t turn my head, actually. So, now, what is good about it is the component technologies–the component technologies of autonomy, the vision, the environmental awareness, the processing, the actuation, the navigation integration, all of those things–those things can be used individually to augment the functionality of the drive. One other aspect is what is called IVHS–Interactive Vehicle-Highway Systems–where the vehicle actually interacts with other vehicles and the highway to decide how to drive itself for maximum efficiency, minimum collision, and so forth. Those are the types of things that would make a material, visible impact for the future of driving and traffic, not so much the autonomy; we’ve had that for a long time. In fact, the old-fashioned autonomous vehicles are even better, because you can hail it, and it shows up at your door, and at the other end you don’t have to worry about parking!
Dan Rubin: Right. What about safety? What are the mechanical aspects of autonomous vehicle safety?
Mark Ehsani: I’m sure you’re aware of this, but you may not have noticed this: Driving is not a mechanical function. Driving is like a conversation. It’s an interaction between human beings. There are issues of culture, habit, psychology, aggression, passivity, timidity. These issues play in. And when you drive behind someone, you immediately realize this person is timid. It could cause certain types of dangers by, you know, unexpected, irrational hard brakes, for example. Or they can be aggressive, and they can pull in front of you, and so on and so forth. And if you go to Mexico, or, in my case, to Spain, to drive, I found everybody was piled up behind me and honking, and I was driving Houston-style, which is pretty aggressive! There’s culture involved. Human beings are interacting and communicating…sometimes with hand gestures and others, but we’ll leave that aside. Can a computer learn to be cultured like a human being. And if it did, how horrible would that be? There are circuits in which autonomous driving is very good. I will give you the first one; I was partially involved in that. The first autonomous vehicle was the Dallas-Fort Worth transit system between the terminals when it was automated, and the driver was taken out. It was a great breakthrough in the early 80s. They’re not on rails, but they’re on prescribed roads. The other one is, for example, stop-and-go buses or delivery vehicles that have a very defined circuit. Those things are perhaps in it. But again, a driver has to be present. So, autonomy doesn’t mean going to sleep, and so on. I end with this: suppose you have an autonomous vehicle truck, and it’s so sophisticated it actually figures out the tempo road, and it rear-ends a small subcompact car, killing a young husband and wife and their two children. Who’s responsible for that? Who are you going to sue? Is there a manslaughter? And whose manslaughter? Is it the designer, the builder, the owner? See, there’s only so much you can penetrate human interaction, because responsibility and safety and cues have to go along.
Dan Rubin: Well, first thing, it’s two things. I thought over what I think of autonomous vehicles and safety or product liability and personal injury or about litigation, potential ramifications, and implications.
Dan Rubin: Right, well, the two things I thought of, when I think about autonomous vehicles and safety, are product liability and personal injury. You know, we’re talking about litigation…potential ramifications and implications. Dr. Pimentel, what about the information technology aspects of autonomous vehicle safety?
Juan Pimentel: It will become more important in the future, particularly now that we’re talking about autonomous vehicles, because, as I pointed out, in autonomous vehicles, we have very big issues, particularly the perception systems. Now, it turns out that for autonomous vehicles the industry is realizing that safety is probably one of the most important things in the design of these vehicles, for fairly obvious reasons I would say. But whenever you’re talking about autonomous vehicles, it turns out that there’s a whole new category of safety, totally different than before. And the reason is that now you’re really having to trust a computer or a system to drive your car. It’s a big difference. So you’re moving away from an actual driver, a human driver, to a computer, to a robot, if you will. Somebody called it robotaxis or robodrivers or whatever. So that is a big issue that I think will have a lot a lot of implications for the law also. But talking about informational aspects, going back to that question, it’s a lot about information technology, IT. In other words, you collect a lot of data from these sensors. I mean, just the amount and the rate of data that you collect is so huge that it’s no comparison to the prior, if you will, technologies or prior stages of automotive [technology]. The amount of data that a radar, a Lidar, or a camera produces is just so humongous. Unbelievable. And the only way that we can process all of that is…even today, we’re having problems processing that. And there are some big players in the industry. You probably just heard one piece of news recently that a company called Nvidia bought another big company called Arm. And one of the reasons for that, by the way, is that they are better positioned to supply the processing power that autonomous vehicles need–not just autonomous vehicles, [also] other things– but autonomous vehicles are a big player in that area. And that’s all related to the question you just asked me: IT technology–processing all of this data–will require massive amounts of processing, massive amounts of new things. We’re just beginning to really have solutions for that. It’s a big problem. It’s a big issue. Very difficult. So definitely a lot of IT aspects, but they’re new. We haven’t dealt with those types of things in the past as you apply them to the autonomous vehicle industry.
Dan Rubin: Do we have any solutions, either short-term or longer-term, to some of these limitations?
Juan Pimentel: Short-term solutions. Let me speak in the area of processing, for instance. The reason I say short-term solutions is that we don’t have a long-term solution yet. The problem is very difficult. Just to give you an example, you probably heard or know that about a year ago, as a matter of fact, Tesla decided to break away from Nvidia, this company that just bought Arm, in terms of using computers for driving the autonomous vehicles. And the reason for that, just to keep a long story short, is that obviously they’re not very happy with the performance of the processors, and Tesla decided to do their own design, decided they could do better. So that shows to you that we have partial solutions, still not long-term, because the problems are very difficult. We’re moving into an area where these processors are multicore. You’re talking about maybe a collection of 20, 30 computers inside a chip, very tiny, smaller than your thumbnail or something. So you’re talking about massive amounts of processing power and memory, all of that into what is called “system on a chip” (SoC). So what we’re doing is combining basically what used to be a full room of computers just a few years ago into a single chip, very tiny. So that’s the level of integration that we’re talking about. So I call that “partial solutions.” We haven’t solved all the problems yet in that there is not just one way of doing that. We don’t have, if you will, standardized or well-accepted architectures for that or protocols. There are some, but not everyone agrees with that. And this is the reason why I don’t see a solution for the next five or 10 years for Level 5 autonomy. These are the elements that will eventually provide you with a long-term solution, but we’re not there yet.
Dan Rubin: Safety issues: in your experience, are they the same everywhere, or do certain countries have different safety issues that are specific to their country or region?
Mark Ehsani: Well, that question I guess has different branches, legally as well as technologically. I think there are differences, yes, but I haven’t really studied that in great deal. I can tell you, I teach this in my graduate course in electric and hybrid vehicles this semester. When I was in New Delhi, I discovered a new vehicle friction which is not in my book, and that is the friction between the body of the car and the body of the people. Because there is such a high density of people, the car is actually rubbing through the pedestrians to get by! So, there, the problem of safety is totally different because the speeds are very, very low, and the injuries are minor and more of an insult than injury. Whereas in the U.S., here, in Austin, near [me], where a Tesla plant is going to go pretty soon, we have a tollway that the speed limit is 85 miles an hour. And, of course, that means people go 90 miles an hour at least. So, that kind of safety is a different kind of safety. It has to do with response time and things like that. I actually developed a technology about 35 years ago, when I was young and reasonably creative, called–not “collision survival technology”, which is what the American cars used to do, big and heavy and basically running through the collision and destroying the obstacle–but “collision avoidance.” The safest vehicles are the vehicles that drive like airplanes. Airplanes are not built to survive a crash. They are designed to avoid a crash, by giving ample warning and then taking over the emergency control from the flyer if they are miscalculating the circumstance of the stall or what have you. And I thought: that technology was actually virtually at hand 35 years ago; primitively, but the elements of it were there. So, those are the types of things I think autonomous vehicle component technologies are really helpful in: collision avoidance, collision mitigation, collision minimization or injury minimization and things like that. So, yes, that technology would be different in New Delhi than it would be in Austin on that tollway, and I think, again, culture has a lot to do with it. In Europe, people are much more aggressive, but the collisions tend to be minor. In America, cars are big, people tend to be more careful, but the collisions tend to be more fatal, with speeds and other cultural traits. Europeans anticipate each other’s faults and flaws. Americans assume the other guy is going to obey the law.
Dan Rubin: Let’s talk about cybersecurity for a second if we can. Obviously, with cybersecurity comes with it the potential for litigation as well. What are the hacking risks when we’re talking about e-vehicles?
Juan Pimentel: First of all, there are too many to mention. But let me answer the question in terms of what are called the main “attack surfaces” in a vehicle. The attack surface basically means the particular areas that hackers can get into a vehicle for attacking purposes from the cybersecurity point of view. Now, it turns out that, on the one hand, we were talking about autonomous driving or electric vehicles that require a lot of new technology. This same technology that’s being introduced for convenience or for automation or for other means, they are the same technology that enables hackers. In other words, they actually add what is called these attack surfaces, particularly the wireless technologies. And the reason is that obviously, with wireless, it doesn’t take any physical means to get into a vehicle and cause all kinds of damage. So, wireless ranges all the way from the well-known phone services, or W-Fi, which is the wireless internet, all the way down to other ones that are not so common; for instance, technologies associated with radar. They are also wireless, if you think about it, because they use the electromagnetic spectrum, just at different frequencies. In fact, it has already been demonstrated that you could hack into a radar system and fool a car to the point that it could cause an accident, just because you hacked into the radar of the car. So, the hacking risks are happening all the time. We were talking about computers just a few minutes ago. It turns out that a new computer architecture, it’s great–and even though today designing a new computer architecture, by the way, is a big issue–it turns out that probably half the effort, maybe not half, but a big portion of the effort, is to talk about how to secure your computer system from hackers. There are many solutions to that. But, unfortunately, in cybersecurity these solutions only last a few years, because the hackers will learn from one another; they publish their findings. And naturally, it’s extremely difficult, even for designers, to make a foolproof design. That just is not possible. In other words, there are vulnerabilities into the system that happen inevitably. And sooner or later, the hackers will get into exploiting these vulnerabilities, so a system could be secure for a number of years and then, that’s the end of it. So it happens all the time.
Mark Ehsani: Fundamentals are stationary as circumstances change. The penetration of computer and information technology in the powerplant control–in other words, propulsion control of the vehicle and vehicle navigation and directional control–those are the vulnerabilities of evil acts or hacking or what have you. And that is the same for electric vehicles as well as conventional vehicles. Electric vehicle does not in any way change the picture. My Jeep Grand Cherokee, it is so computerized that…in fact, there were lots of articles, you might have seen it, about hacking into the engine control and the brake control and so on and so forth, and somebody actually demonstrated that that can be done external to the vehicle. And actually I had a case of that, which I feared was hacking, but it t urned out to be not: the car wouldn’t shift gears, and I had to travel 230 miles virtually in first gear to get to the repair shop. So that’s the first thing. The second thing is–and here I’m skating on thin ice, but it’s my observation–hacking takes two elements: it takes a computer vulnerability, and it also takes a fool. You have to have a fool that lets the hacker in. Sometimes it’s called “foolish updating” because everybody wants to have them. Sometimes bells and whistles and favors, or false threats, or what have you. Most of the great hackings, in stationary and mobile experience, have taken more than a computer vulnerability. It has taken a fool that can be persuaded. Again, it’s a human interaction issue. Hackers will never be removed because there is one born every day or every hour. Even if you perfect the computers.
Dan Rubin: Sure. I read recently that the charging stations for these e-vehicles are at risk of hacking. It makes me think of every time you go to an ATM and put your card in and how, obviously, there is that potential there. So maybe I’m the fool.
Mark Ehsani: That’s why I don’t use the hotel internet hotspot… because it’s so convenient for fools!
Dan Rubin: Right. Right. Exactly. I’m going to make note to not connect to the public networks anymore. So, wanted to open it up to questions from our participants. Looks like we have one from Brice David. I’m going to unmute Brice so he can ask.
Brice David: Thank you. I’ve had luck in my career to work with both of these gentlemen. With Dan’s caveat that you’re not attorneys and you’re just engineers, and we talked a little bit about litigation that might happen around the battery area because that’s one big area, I’d like to see where else, where else are we seeing innovations coming, either autonomous vehicles or e-vehicles, that are so important to the core of making it work that it’s likely to be the areas where people are going to be fighting, basically…litigating?
Mark Ehsani: Mine is easy, so let me go first. Electric propulsion is extremely safe, reliable, and can be multi-layer fortified against failures. That’s not an issue. It’s all about battery vulnerability, and it’s something that is beginning to be better and better understood. I don’t think that’s going to be an open-ended or progressively worsening problem. So, I don’t think electrification of vehicles, including their battery technology, is going to be a source of additional failures or vulnerabilities. I think, if anything, it is going to be better because you don’t carry a paper-thin gas tank that can rupture upon impact, and not to mention leak. So really the risk is all Juan’s. It is all about introducing autonomy, intelligence, trying to displace or compete with the human driver.
Juan Pimentel: Yeah, my point on that is what I mentioned before also: Level 3 that I mentioned before, it opens up a lot of issues…sorry, could open up a lot of legal issues…because this is an area where the driver is involved partially. And obviously, the driver could misuse the equipment or could make mistakes, and same thing for the automation…could make mistakes. So we see Level 3, even higher levels, but Level 3 particularly, in my opinion, is going to open up a lot of legal models. That’s where I expect, perhaps not most, but a lot of lawsuits. As a matter of fact, some of the OEMs, believe it or not, they have avoided–I’m not sure if they are still doing that, working on Level 3–they wanted to jump all the way from 2 to 4, precisely for that reason. There are just too many issues. And I think they wanted to avoid lawsuits for the most part. But, having said that, I don’t believe they can avoid that. It turns out that the industry today is heavy into Level 3. I don’t think you can avoid it…I mean jumping Level 3. The reason is that it would be too big of a jump going from 2 to 4. You have to go gradually, and I think the industry already decided that it’s such a big problem that you have to go, whether you like it or not, to Level 3.
Dan Rubin: Hmm. Well, thank you, Dr. Pimentel, and thank you, Professor Ehsani. We’ve reached the end of the hour. I think that’s a perfect question to end with. Thank you, Brice. I want to thank everyone who attended today for taking time out of your busy schedules. But a special thank you to Juan Pimentel and Mark Ehsani for sharing a part of your day and your expertise with us. Really appreciate it. Thank you again. And have a great rest of your day, everyone.
Speaking to experts helps craft good strategies and tactics. If you are an attorney and are interested in discussing your own client’s specific situation with either Professor Ehsani or Dr. Pimentel, please contact Dan Rubin: firstname.lastname@example.org or (202) 908-4492.
After a quarter century helping litigators find the right expert witnesses, Round Table Group’s network contains some of the world’s greatest experts. On the Discussions at the Round Table podcast, we talk to some of them about what’s new in their field of study and their experience as expert witnesses.
Dr. Juan Pimentel, automotive industry consultant is an expert in the areas of autonomous vehicles, automotive in-vehicle networks, automotive safety and cybersecurity. He has published extensively and has written several books. One of his latest publications is a series of five edited books for SAE International on various topics related to the safety of automated vehicles. He is on several ISO and SAE committees involving safety of the intended functionality (SOTIF) and vehicle cybersecurity. He is an experienced expert witness.
Professor Mark Ehsani is the Professor of Electrical Engineering and Director of Advanced Vehicle Systems Research Program and the Power Electronics and Motor Drives Laboratory at Texas A&M University. He is the author of over 400 publications, 20 books, and over 24 US and EU patents in pulsed-power supplies, high-voltage engineering, power electronics, motor drives, and advanced vehicle systems. Professor Ehsani has worked on over 50 matters, many of which were patent-related.
2.7% of the U.S. gross domestic product can be attributed to the automotive industry, which employs about 17.9 million individuals. Our automotive expert witnesses, speakers, and consultants include scholars, mechanical and chemical engineers, researchers, inventors, and veterans of the automotive industry with knowledge in a broad range of topics.