On Tech Ethics Podcast – Technological and Safety Considerations for Autonomous Vehicles

Daniel Smith: Welcome to On Tech Ethics with CITI Program. Our guest today is Patrick English, who is a professor at Ferris State University School of Automotive and Heavy Equipment. Today we are going to discuss technological and safety considerations for autonomous vehicles.

Before we get started, I want to quickly note that this podcast is for educational purposes only. It is not designed to provide legal advice or legal guidance. You should consult with your organization’s attorneys if you have questions or concerns about the relevant laws and regulations that may be discussed in this podcast. In addition, the views expressed in this podcast are solely those of our guests.

And on that note, welcome to the podcast, Patrick.

Patrick English: Thank you very much.

Daniel Smith: Now tell us more about yourself and what you currently focus on at Ferris State.

Patrick English: Well, I teach in the Automotive Engineering Technology portion of our program, which is a four year hands-on vehicle engineering degree. I focus primarily on alternative fuels, hybrid powertrains, essentially trying to get the latest of what’s just around the corner in the automotive industry to our students before they graduate.

Daniel Smith: Something that has been in the news a lot recently is the progress that is being made on autonomous vehicles. Can you start by providing an overview of the different types of autonomous vehicles?

Patrick English: Generally, the autonomous vehicle categories are broken into five separate categories. The first one is a Level 1 because in Level 0 it’s just your regular old car where you do everything for it. Level 1, it’s got some advanced driver systems and the vehicle can sometimes assist the driver with either steering or braking or some acceleration, but not simultaneously, not altogether. Level 2 vehicle can control both steering and braking simultaneously under certain circumstances, but the human driver needs to pay full attention to what’s going on and the surroundings and the situation the vehicle’s in. Level 3, vehicle can perform all the aspects of the driving task, but some circumstances it’ll get into situations where it just does not have all the data that it needs to completely take the place of the driver. Level 4 vehicle can drive itself, perform the driving task, monitor the driving environment, and do basically all the driving, and for the majority of the time, the driver doesn’t have to pay a whole lot of attention to what’s going on unless there’s some kind of an alert from the vehicle. Level 5 is our fully automated driving system, where the vehicle does all the driving in all the circumstances. In some of those vehicles there are some designs that, say General Motors, with cruise have out there that don’t even have a steering wheel. There’s no opportunity for a person to take over what that vehicle’s doing.

The level 5, I think that’s not a reality for open road right now. They have some technology that’s pretty close, but generally they’re testing it with a person in the vehicle so that in case it does get to a point where data is not complete for the vehicle, the person can take over. Generally right now what’s out there is primarily Level 3. You still have to be the driver under certain circumstances and certainly it’s in your best interest to continue to pay attention to what’s going on around you to make sure that there isn’t some area of the data that the vehicle cannot interpret.

But when it comes to Level 5, there is some of that out there, but it’s not open road. It’s fixed course, so say like a shuttle bus on a college campus or something like that, that is a good application for a Level 5 autonomous vehicle or bus or something along those lines because you can fix the variables a little bit more. You can anticipate a little bit more in terms of the variables because it’s always going down the same path, it’s always on the same track, whereas open road, there’s going to be a lot of road conditions, weather conditions, and things that are a bit more challenging for data interpretation.

Daniel Smith: Now how close are we until Level 5 vehicles are on the open roadways?

Patrick English: I don’t think it’s necessarily a matter of the technology as much as it is the cost to the technology. The LIDAR suite, which is one of the setups that they’re using to do this data gathering, the cost is still prohibitive for your average four door sedan, it’s going to put the cost up rather high. Also the energy level of all of this data acquisition and the sheer volume of information that the vehicle’s processing, it requires quite a bit of energy. In general, I think it’s going to be currently the cost that’s going to make them take a little bit longer to get out on the road. At last count, there are 26 companies that are working on their own autonomous so it’s certainly not for lack of research, it just right now the technology is costly to the point where it’s not going to be an everyday thing until the cost comes down or until they find a way to do it and use less energy or less expensive sensors.

Daniel Smith: Aside from cost, what are some of the unique ethical and safety considerations for autonomous vehicles in general and in Level 5 vehicles?

Patrick English: Well, in Level 5, like I said, right now, they’re going to be primarily fixed track things. There are some fully autonomous buses that go on fixed routes, which with the quality of the sensors and their ability to stop the vehicle, if something’s in front of it, they are pretty safe, but they’re not high-speed applications by any means. That speed limits in the country range anywhere from 55 in areas that are still federal to 85 in areas that are not under the federally mandated speed limit, or what used to be federally mandated speed limit, but those areas still exist, so anywhere between 55 and 85, and of course increased speed means you’ve got to process things that much faster. In those areas, of course you have high rates of speed, that’s going to change really at what point a Level 5 is going to be realistic.

There’s a lot of interesting things that they want to do with that level of technology like vehicle trains to essentially speed up transportation for people. With Level 5, I think there are a lot of really good positives once we do actualize it, actually get to that point, especially folks who are no longer driving and need transportation and things like that. This could be a big deal, a life changer for a lot of those folks but I think it’s going to be a while before we see full Level 5.

But as we look at some of the implications that are out there, Level 5 vehicle gets into an accident, who’s responsible? The person’s going to say, “I was in the car but I wasn’t driving.” It’s been a few years but one of the things that the computer systems had a terrible time with for a while was people on bicycles, spinning wheels or something about the spokes and the wheels or just the data picture was not good and they were having trouble identifying that. Or there was a case with a self-driving vehicle that turned into another vehicle because the side of the vehicle, it was a semi, had a mirrored surface on it and the vehicle’s computer couldn’t really compute what that was and thought that it was open space. In these instances, if it is in fully autonomous mode, who’s responsible for it? We’re dealing with millions of people, so lots of interesting situations are going to occur. Somebody gets behind the wheel and they’ve been drinking, well, they put it on fully autonomous mode, are they driving while intoxicated? It raises a lot of interesting questions. What are our responsibilities and where does that fall?

Unfortunately the guidance isn’t there yet. July, I think’s the last time I looked, there were I believe 29 states that had not done anything with regulations concerning self-driving vehicles. There are no federal laws, there are guidelines from the NTSA but they’re strictly guidelines, they’re not hard and fast rules. To use an expression, we’re still a little bit Wild West when it comes to autonomous vehicles.

Now, Society of Automotive Engineers is at least classifying things and they have a number of committees and things that are looking into what are the issues that we have to deal with and what’s going to happen with these vehicles, but we’ve got a ways to go yet before we have a set of rules for them. Another interesting thing to think about is why are we looking at autonomous vehicles?

Well, for one thing, people are people. Everybody’s trying to multitask and we’re dealing with distracted driving, which is one of the largest causes of injury today in young people, and there’s some statistical data to say that 76% of all accidents are based solely on human error and in 94% of cases, some human error of some sort is involved. Based on that data, and I certainly would encourage you to look into the data more closely because it’s always good to look into the background of the studies before you put too much emphasis on them, but if that is truly the case, then there’s going to be an interesting thing happening with insurance because it’s common that a self-driving vehicle may cost more to insure currently whereas in the future, if the major push is for self-driving vehicles, what’s to say that part of that push isn’t done through insurance where if you choose to drive, then they’re looking at this data and they’re saying, “Well, you’re more dangerous because you’re the person behind the wheel,” and that’s going to make your insurance expensive enough that driving yourself may become cost prohibitive.

Daniel Smith: That is interesting, and in a moment, I want to get back to some of the measures that manufacturers and others are taking to address some of the issues that you brought up, but first I want to take a quick break to tell you all about CITI Program safety-related content. Our environmental health and safety courses provide expert-developed training on topics such as laser safety, radiation safety, and laboratory chemical safety. You can learn more about these courses and others at citiprogram.org. Now back to our conversation with Patrick.

Now I want to touch on the continuous improvement of autonomous vehicles and the role of real world data collection. What role does real world data collection play in improving the safety of autonomous vehicles?

Patrick English: It plays a huge role, and they’re continuing to do live studies of these vehicles all over the country. You can go in many of the urban areas, I was recently in Ann Arbor and there were autonomous vehicles driving around that were test vehicles, so they did still have a person in them to make sure that if it got into some kind of a situation that the person could take over, but they were driving the streets of Ann Arbor and collecting data in how to handle situations with the vehicle. I was in Vegas a few months ago and a similar fleet of autonomous vehicles were driving in Vegas and collecting data. There’s quite a few companies out there that are really trying to quote unquote, “Crack this nut.” I think between the different types of sensors and the different approaches to it, I think eventually with the volume of real world data, they’re going to be able to come up with something where it is going to be a pretty solid system.

Not to mention as we get into vehicle to vehicle communication, the vehicle 100 yards or 200 yards or even half a mile ahead of you may encounter a situation that it has to interpret, icy spot on the road, whatever. Once it interprets that situation, once it handles that situation, it can then send that information out to the vehicles around it so by the time you get to that icy spot in the road, your car has already prepared to make sure that it remains stable across that icy spot on the road because it can actually use the real world data to say, “Okay, I know that there’s ice over here. I know that there’s not going to be appropriate traction on this side, so have to operate the powertrain accordingly so that we don’t slide.”

Daniel Smith: In addition to data collection, are there other measures that car manufacturers and others are taking to address safety issues?

Patrick English: I think part of it is being cautious as to beginning with your Level 3 stuff and making it so that it may be more capable than Level 3, but you’re expected to operate it as Level 3 and not push beyond, such as the Cadillac system where if you take your eyes off the road, the vehicle can begin to slow down, the vehicle can take action with a haptic warning or whatever to essentially get you to, after so many seconds. Put your focus back on the road. Even though it has a good Level 3 system, they have ways to make sure they’re monitoring what the driver’s doing and they’re going to make sure that the driver still maintains their attention with the road.

Of course, all of this happens while the vehicle’s still collecting data, while they’re still perfecting what they’re doing. The progress is going to continue while these vehicles are in the road. One of the good things about what the manufacturers are doing is that they are going to continue to actively collect data. One of the early things that OnStar really helped with is that the manufacturer was able to, on certain vehicles that had a lot more research possibilities, hybrid and electric vehicles, those connected technologies allowed the manufacturer in many cases to pull more data from the vehicle and make the next incarnation of that vehicle even better.

Daniel Smith: Then on a regulatory level, I know that you mentioned that there is not a federal law or regulation governing autonomous vehicles specifically, and that a lot of states are a bit behind the curve here, but it sounds like there are some states that regulate autonomous vehicles. How do some of those regulatory frameworks that are in place work?

Patrick English: Well, it’s very mixed. There aren’t any states currently where it’s illegal to operate an autonomous vehicle, but there are some regulations, and I don’t have an exhaustive list, but there are some regulations where the expectation is still that the driver will be in condition to operate the vehicle and paying attention to their surroundings. Some of the states have put their autonomous vehicle regulations to a vote and it’s kind of come up with a mixed result. They haven’t necessarily gotten all the different things that we’re looking for. It’s not illegal and there are varying levels across the states currently.

Daniel Smith: You also mentioned some industry standards earlier. Can you tell us some more about how those come into play?

Patrick English: The industry standards are more along the lines of things that the manufacturers are putting out there as information with their vehicles. The most important thing I think for them is to set up a realistic expectation for the customer so that they don’t essentially have greater expectations of the system than it is safe to operate it at. The majority of it is just that they’re trying to lay out what they believe is the safest and also how far you can use the system safely without asking more of it than it’s going to be capable of.

Daniel Smith: Are there any additional resources our listeners can check out to learn more about autonomous vehicles, safety issues, regulations, and other related topics?

Patrick English: There are a number of different good think tank groups that are really talking about what’s going on with autonomous vehicles. In terms of regulation, there’s a white paper by the firm called Jones Day that has a lot of really good data in it. Certainly SAE has regular meetings on it, and you can get on the SAE website and there’s a fair number of interesting things that you can find on there. Really realistically, it’s a hot topic on the internet and there is quite a bit of solid data that can be gotten.

Daniel Smith: Wonderful. I will certainly include links to these resources in the show notes so our listeners can learn more. On that note, do you have any final thoughts you would like to share that we did not already touch on?

Patrick English: Well, I think it’s interesting all the changes in the automotive industry. It’s an exciting time to be teaching advanced technology in the automotive world. It’s certainly an exciting time to be teaching at Ferris where we really make sure that our students go out from here and find some really interesting and lucrative places to work in the automotive world and with the automotive manufacturers and the Tier 1 suppliers.

Another thing I found interesting, and I wanted to mention it just because it’s something that I’m not sure most people are aware of, but there is a remote driving vehicle that is out there. Essentially what happens is that the vehicles are autonomous and if you get into a situation where whatever’s going on is beyond the capabilities of the vehicle’s processor, it gets into a situation where there’s some ambiguity, there’s something that it can’t figure out, it will actually switch control of the vehicle to a person at a, basically it’s a call center where they have multiple screens set up and the control of the vehicle goes to that person and their computer becomes whatever vehicle you’re in, and they guide the vehicle through whatever portion of the driving that the vehicle was unable to do. There is a remote option that they’re trying out on some of the autonomous vehicles just to take care of those situations where the data isn’t enough or the data isn’t able to be made sense of. I think that’s something interesting that a lot of people aren’t aware of.

I was fortunate to be in one of these vehicles to really get a chance to see what it could do, and they simulated a situation where the vehicle was not able to handle whatever was coming in terms of data and it was interesting to be there where it switches over to a computer console somewhere else and the vehicle continued to drive.

Daniel Smith: It will be really interesting to see how that technology evolves and plays into the safety of autonomous vehicles.

That is all for our conversation today, so thank you again, Patrick. I also invite everyone to visit citiprogram.org to learn more about our courses and webinars on research, ethics, and compliance. You may be interested in our Privacy and Ethical Considerations for Connected and Automated Vehicles webinar, which discusses the basics of privacy, data protection, and ethics in the context of connected and automated vehicles. With that, I look forward to bringing you all more conversations on all things tech ethics.