This week’s podcast: Automakers are cramming new safety features into their vehicles. What’s it like actually driving with all of that stuff? A free-ranging discussion with automotive writer and consultant John Voelcker on driving poorly on purpose, weird algorithms, and the pitfalls of an entire industry possibly leaving its customers behind.

[FULL TRANSCRIPT WILL BE AVAILABLE SOON]

BRIAN SANTO: I’m Brian Santo, EE Times editor-in-chief. You’re listening to EE Times On Air, and this is your Weekly Briefing for the week ending April 16th.

EE Times has been closely following the development of all the technologies the auto industry is adopting in their vehicles as car makers attempt to improve the experience of driving: everything from the entertainment systems in the cabin, to the advanced sensor technologies to improve safety, to the artificial intelligence systems that can already drive on their own in strictly controlled circumstances.

What we haven’t done much of is actually get in cars and drive them around to find out how all this whiz-bang technology actually performs. This week, our guest is John Voelcker, an automotive reporter and consultant who’s been test driving cars for pretty much the entire history of assisted and autonomous driving. Here’s JV on one of those earlier efforts:

JOHN VOELCKER: I always tell the story about Volvo, which had one of the very first automated emergency braking systems, which is another in this constellation of ADAS systems. Volvo’s engineers very earnestly explained to us journalists that they were very worried that people would come to rely on automatic braking. They deliberately made it absolutely as unpleasant, violent and scary for the driver as possible. I experienced that, in fact.

It’s very hard for any driver, especially for someone like me, to drive a car at an object and not brake.

BRIAN SANTO: We’ll get back to my conversation with JV in a moment.

First, here’s a quick run-down of some of the top articles we have in EE Times this week.

*Artificial intelligence is already seeping into more and more of our everyday lives, and it’s only going to get more pervasive. That makes investors incredibly excited about getting their money into AI. We’ve got stories this week about humungous funding rounds for two prominent AI companies: SambaNova, already a unicorn (uh, company) perceived to be worth at least $1 billion, just went quintuple unicorn. Meanwhile, Groq had so many investors trying to give it money, it had to tell a few of them to take a hike. The problems you’d like to have, right? We’ve got one story on each company, what they’re doing, and what they expect to do with the money they just got.

*Graphics processor specialist Nvidia announced its first CPU, and it also laid out its plans to further permeate the automotive electronics market. We’ve got coverage on the web site of those two developments, but Nvidia actually announced way more than that. We’re planning on going over the whole kit-and-kaboodle on next week’s podcast, so there’s that to look forward to.

*We also have stories on a satellite rescue in space, Roku’s new operating system that makes watching TV easier, the US military’s contract with Microsoft for augmented reality technology, and a nifty new copper via technology for printed circuit boards, which are getting loaded and packed into ever-smaller, tighter spaces, and then still have to deal with thermal dissipation.

You can find these and other stories when you visit our web site at eetimes.com. If you reached this episode through our podcast web page, there are links to all of these stories on your left.

The automotive industry is in the midst of an amazing transformation. The internal combustion engine is gradually being eclipsed by electric power. Dashboards are becoming computer displays, infotainment systems are rivaling home video and audio equipment, safety features are becoming ever-more sophisticated.

We’ve been closely following the development of the technologies for all of this stuff — battery technology and lidars and software stacks and, importantly, how it all could lead to greater driving safety. But what does it all add up to in terms of driver experience?

We just happen to have a friend who test drives cars for living. John Voelcker has been writing about technology for a long time. He’s a car enthusiast, and somewhere along his career, he figured out how to merge his profession with his passion, and he became an automotive writer and consultant. For nine years he ran Green Car Reports, which covers the automotive industry with a specific focus on electric vehicles and hybrids.

Lately he’s been working as an independent writer and consultant, operating out of a secret lair somewhere in the Hudson Valley. Here’s our conversation.

BRIAN SANTO: When did you start getting to test drive vehicles that had these types of features and how did they develop, hopefully progress, over the years? The first ones were things like rear-facing cameras for help backing up. That sort of thing?

JOHN VOELCKER: I tend to think of the first automated driver assistance function as cruise control, really. Plain old fashioned cruise control where you could set it, you could take your foot off the accelerator, but you had to watch because the car wasn’t aware of what was going on around it and it would quite happily drive you into a stopped car up ahead.

I’ve been test driving cars for now for I guess 12 years. Back when I started in the late 2000s, it was unusual to get a car with adaptive cruise control. I believe Mercedes Benz developed that and it was called Distronic. (But fact check me on that.) But it was still a relatively high-end thing. And over that decade plus, what’s been interesting is seeing these fairly complex and expensive systems sort of migrate the vast majority of the autos that are available today. Not necessarily as standard features.

The word to look for is “available,” which means extra-cost option in autospeak. Even compact sedans of the sort that used to be called econoboxes, many of them now come with a suite of ADAS features that would have been a very expensive option even 10 years ago on a premium car.

Once you had adaptive cruise, that was the first time the car really had a sensor to think about what’s going on around it, and that was a huge advance. We forget now.

People like me who test drive cars, we tend to get all of the press leads. They just put all of the top-end, most lavishly featured vehicles into the press lead. So we get to test them with all the bells and whistles.

Ten or 12 years ago, getting a car with adaptive cruise was exciting! Here’s this new technology that I can test. And then you had the evolution of sort of land awareness.

There are a number of pieces of that. First, it was really just lane departure warning. And however the car did it, it essentially looked at the lines on the side of the road, and if you were drifting, if there was no steering input and you were about to cross a line, it would alert the driver. I found it interesting: There’s actually two ways to do that. One is beeping, of course. There are a lot of beeps for a lot of these systems.

Certain carmakers said, Well, drivers react badly to beeping because other people in the car know that they screwed up. So what we’re going to do is vibrate a corner of the driver’s seat on the side that’s departing. You really feel it. There might be a little purr or rumble that the passengers heard, but not if the radio was up. That way, the driver knows that he or she is drifting out of the lane, but the rest of the passengers don’t. Which I thought was interesting. Nonetheless, we still have a lot more beeps today than we do vibrations.

So lane departure — just sort of, “Hey, you’re drifting out of your lane” — was the first step. Then you had lane departure warning, LDW. Then you had automated lane correction, which essentially not only said, “Hey, you’re about to drift out,” but steered you back away from the lane line. That was the state-of-the-art sort of in the mid-2010s. That was where it started to get interesting because people like me, who find a nice MP structured freeway, set up the car, put it into lane-correction mode, and keep our hands lightly on the wheel and see what happened.

There was really a difference between systems. Some of them would correct and then stay roughly straight. The worst of the systems really ping-ponged you between lane lines. I actually made myself slightly queasy once in a mass market vehicle whose lane correction was fairly sudden and fairly abrupt compared to some of the others that were more gentle. Literally, I ping-ponged it for two or three miles, and it really was an unpleasant feeling.

That gets us to this notion of algorithms. One of the things that has probably driven these systems from 12 or 15 major global makers. While the screens will be different and the controls will be different, software underlying, let us say, a Toyota ADAS system is pretty much the same as the software underlying a Lexus ADAS system, because Lexus is Toyota’s luxury brand.

Watching the evolution of these algorithms — and some of the rough edges got smoothed off over time — but finally, after lane correction, you had what is called active lane control, which basically adds lane centering. Not only does the car notice you’re drifting, not only does it correct for that drift, but it essentially figures out where the center of the lane is and puts you there. You still have to keep your hands on the wheel, but it will essentially do your cornering for you. If the corners are relatively mild and predictable and if you’re not doing anything like passing another car.

So if you combine that with adaptive cruise control, then all of a sudden, you have most of the elements for the first real level of autonomous driving.

BRIAN SANTO: Do you consider that Level 3 or Level 4? Our listeners will probably understand the distinction.

JOHN VOELCKER: Level 3. There is a lot of automaker discussion about levels that gets slightly choppy around the edges.

BRIAN SANTO: Yes.

JOHN VOELCKER: Sometimes I’m not entirely sure the loop gets 100% closed between the people who write the copy, the folks who turn it into press releases and the actual engineers who are designing this stuff and actually read SAE definitions.

For that reason, honestly, because I write to consumers, I stay away from levels. That may not be terribly helpful for you, but what I try to do is describe what the system actually does for the driver.

BRIAN SANTO: We recently ran an article where we quoted one particular automotive expert, a fellow named Phil Koopman, who said that the levels don’t translate well to consumers. Stop talking about them. And he proposed a set of categories where the point isn’t what the car does, the point is what the driver’s responsibility is at each level.

JOHN VOELCKER: Oh, I like that! What the driver has to do.

BRIAN SANTO: Right.

JOHN VOELCKER: Honestly, the GM Super Cruise system, the best definition I’ve seen of it is adaptive cruise lets you keep your foot off the pedal. The Super Cruise system, which has driver monitoring built in (so lane centering and essentially keeping you in your lane). The Super Cruise system is not only feet-off but hands-off. I thought that was a very nice, simple, graspable way to explain what it did and didn’t do and what the driver did and didn’t have to do.

BRIAN SANTO: With your experience with the vehicles you’ve seen and that you understand are coming up, are we still talking about highway driving? Or does this include city driving?

JOHN VOELCKER: My suspicion is that autonomous or self-driving vehicles (both of them being terms that I have problems with, which I’ll get to in a sec)… My suspicion is that there will be two parallel channels. Highway driving, at least in North America, is relatively easy most of the time because interstate highways are built to a very specific and predictable set of driving criteria. And so you’re not likely to get all of a sudden a stop sign hidden by a hill because the interstate highway design specs have rules about vision.

I think most North Americans will first experience these things on the highway essentially as an expansion of the world that cruise control lives in. But I think there’s a separate category of self-driving vehicles that may not be privately owned that operate in very carefully geofenced and intensively mapped districts. GM’s cruise unit is mapping San Francisco.

San Francisco is a tough place to drive. If you can do a self-driving vehicle in San Francisco, you’ve covered a lot of edge cases. The thing about urban driving, though, is that let’s say you just have a blanket speed limit in San Francisco at 30 miles an hour. That makes it much easier to be safe because a modern vehicle, at the limits of its performance envelope, can stop you really suddenly and really fast from 30 miles an hour. But physics comes into play, and it’s much harder at 75 miles an hour.

BRIAN SANTO: Right.

JOHN VOELCKER: I always tell the story about Volvo, which had one of the very first automated emergency braking systems, which is another in this constellation of ADAS systems. Volvo’s engineers very earnestly explained to us journalists that they were very worried that people would come to rely on automatic braking. They deliberately made it absolutely as unpleasant, violent and scary for the driver as possible. I experienced that, in fact.

It’s very hard for any driver, especially for someone like me, to drive a car at an object and not brake. But I beat down my subconscious, did that, and all of a sudden a bright red light at the base of the dash started flashing. And a second or so later, a really, really unpleasant buzzer — kind of like the nuclear plant alert buzzer — started going. And then, when I continued not to brake, the car slammed on its brakes absolutely as hard as possible, throwing me into the seatbelt to such a degree that I probably had a bruise. And I will tell you: I never wanted to rely on that system because it was really unpleasant! Which I think is very smart. Because god knows, distraction (mobile phones, texting, kids today… waving my rake). Distraction is a huge problem.

One of the worries about some of these systems is that they are simply amplifying what are already the encouragements based on the odds to not pay attention to driving.

I’ve experienced GM’s Super Cruise several times over a period of about three years now and watched it evolve. And it watches you. You can take your hands off if the car isn’t operating in an area that has been mapped to a very great degree of detail, and all the sensors and working, and you’re buckled in and your passengers are buckled in and so on and so forth. And it works very nicely. And it mostly drives as you would, plus or minus.

But if your eyes drift for more than a few seconds, it will flag you.

One journalist, I forget who it was, came up with the question: So you’re on Super Cruise; what happens if the driver dies? The answer is: GM had an answer for it. Some engineer said, Oh, we thought about after a certain degree of no driver input — whether the eyes aren’t pointing in the right place or there isn’t any steering input or whatever — the car will turn on its flashers, slow itself down and safely park itself at the side of the road and call 911.

BRIAN SANTO: Wow!

JOHN VOELCKER: Yeah.

BRIAN SANTO: Have you ever experienced… I’m not going to ask you if you’ve ever done this, but are you aware of anybody who tried to deliberately fool Super Cruise wearing dark glasses or deliberately riding at some time of day when the sun was beating directly into the cab? Anything like that?

JOHN VOELCKER: I have not. When I test cars, it’s for a whole variety of things. One of them are the automated safety systems. But I’m not a specialist in that area, so I have sort of a limited window.

BRIAN SANTO: Yeah.

JOHN VOELCKER: I will say, subjectively, because I do a lot of driving, I’m always somewhat surprised at the degree to which the different makers’ algorithms really do produce a different automated or semi-automated driving experience.

Tesla, I would venture to say, is one of the top two or three in providing what I would say is a natural driving feel. A lot of it is really subtle stuff, especially for people who drive a lot.

I had a long discussion with a Nissan engineer when I went to their test track several years ago to drive the new version of one of their systems. One of the things it turned out not to do, which is a standard human behavior, is if you’re right next to a semi and you’re in a much smaller vehicle, you will unconsciously drift toward the outside of your lane to put some extra distance between you and something that is twice as high as you are and four times as long and weighs 60,000 tons. Most people aren’t aware of doing it, and I wasn’t aware of it until the automated Nissan did not do that. This was probably 2015 or ’16. And I suddenly realized that it made me extremely nervous to be operating 15 inches from the side of a semi.

The car was keeping itself in its lane, and it would notice if the semi drifted over its lane. But there is a lot of subtlety to how people drive. Mostly these algorithms aren’t quite there yet.

Another thing is that people clip the insides of turns and drift wide on the outside.

BRIAN SANTO: Guilty!

JOHN VOELCKER: It’s a version of the racing line, basically. I have found myself fighting with some of these systems because I will go toward the inside of the curve, and the system says, “No. You’re not on the center line. Stay where you are.” There’s a lot of subtlety to this stuff.

That said, 90-x% of the time, they work fine. Part of my standard test loop is about 80 miles on an interstate highway. Just to see how the car does, how all of those systems work, and essentially for those few that have something like Super Cruise (and Ford actually just announced its own sort of take on hands-off driving this morning, which will arrive first in the Mach-E electric Mustang, and of course their best-selling F-150 pickup truck). Those systems, there’s a lot more to test to see how long you can look away and so on.

After a while, essentially active lane control and adaptive cruise do make things a lot calmer and quieter, especially if they’re in an electric car, because you don’t have all the shifting up and down and so forth. But there are differences in feel.

The challenge for me is that I drive all these cars all the time. Someone who test drives a car may not even test out these systems to see if they like the feel, much like do it for three different cars that they’re considering buying. It’s unfortunately not how dealerships and test drives work.

BRIAN SANTO: As you mentioned, you’ve been doing this for about a dozen years and getting introduced to these technologies as they’re introduced and experiencing them as they get developed. And you’re writing for consumers. You have a weird experience in that you have a continuum of experience and you jump from car to car to car. The average consumer might be going from like their 2009 Jetta or whatever that may have a few things and now all of a sudden they’re dumped into this ultramodern cab that has this incredible suite of features. Do you feel that there’s a learning curve? If there is a learning curve, do the automakers seem to be aware of it?

JOHN VOELCKER: There is definitely a learning curve. Automakers are aware of it and relatively uncomfortable, because this is why I point out that, under the laws of almost any state of the 50 in the US, automakers are not legally allowed to sell you a car. You must buy a car from an independently owned and operated third-party distributor called a car dealer. And automakers have a variety of ways of influencing the behavior of those dealers that sell their products to consumers, but they cannot control everything they do.

It has been clear for the last 10 years that, in some cases, the pre-delivery explanations when someone comes into one of these new cars… many people are driving cars that do not have center screens, that do not have the ability to pair your smartphone, let alone automated driving assistance systems of a whole variety of flavors.

BRIAN SANTO: That’s me!

JOHN VOELCKER: When I’m not test driving cars, our runner is a 2000 Subaru Outback. Which has cruise control, which is an enormous advance over the Subaru before it!

One of the major and continuing challenges is trying to get dealers, once they have made the sale, to do a proper briefing once the car is delivered. And to explain a lot of the stuff, frankly, takes 30 to 45 minutes. And most consumers are not going to retain most of it.

That said, most consumers do not use most of the electronic features that are built into their cars. Most consumers will probably set up favorite radio stations (if they listen to terrestrial radio or Sirius XM), they may pair their phones (especially those who spend a lot of time driving), they may play music off them, and they may figure out how to make the cruise work. But my gut, just from talking to a lot of consumers and automakers who are very cagey about that stuff (but that’s why god made alcohol)… The good journalist is always one drink behind the source.

BRIAN SANTO: God bless you.

JOHN VOELCKER: I get to say stuff like this these days.

It is a problem. Remember also the structure of salespeople. The salesperson is only making money when he or she is selling. Spending 45 minutes explaining the $40,000 you just sold to the person who wrote you a check last week is not making you money. So dealers have different ways of doing that. Sometimes they have delivery teams.

The automakers put a lot of effort into training those delivery teams because their consumer satisfaction scores suffer from the lack of those teams. People will down-rate their cars because they say, “It’s hard to put it in cruise. My last car had a lever to get to cruise, and now I have to turn it on with a button on the wheel, and then I have to press a different button to set it, and then I have to do a third button to make it speed up or slow down, and then there’s this other thing where all of a sudden it’s adaptive, which it never was before.” And that’s just cruise control!

Cars have gotten very complex very quickly. Before we even get into the idea that you can customize your car’s drive modes and suspension settings. You can pick what color lights you want in the interior. You can pick whether you want a tachometer or whatever. There are enormous options in a car, most of which most drivers will never use. It is a continuing problem.

BRIAN SANTO: A problem because people don’t use them at all? Or when they kick in they don’t expect them and get surprised, and that’s a problem driving?

JOHN VOELCKER: All of the above. All, unfortunately, means that the car doesn’t behave the way the last one did, and it startled me or I couldn’t figure it out, so the car’s not as good as the last one.

BRIAN SANTO: Here’s an interesting thing: We talk about how all the different technology companies that are supplying the automakers are different and of course all the automakers are different. I’m imagining what happens if you’ve got a family that’s got two or three cars and the family members switch off. Switching between two cars whose safety packages behave in different ways has also got to be a concern.

JOHN VOELCKER: I would think so. I think probably two you can toggle. But if you have multiple vehicles in the household, then it gets different. Honestly, I have gotten into test cars, set the cruise, I’m focusing on accessing something else and then I realize, “Oh! This is not adaptive! Hit the brakes!!”

BRIAN SANTO: Wow! Yeah, I can see where that would be alarming.

JOHN VOELCKER: That’s sort of the other end of it, expecting too much. It speaks to the very great difficulty that automakers and their marketers have explaining technology to shoppers and buyers who are only sort of paying attention. Until that three-week window where they are going to buy a car.

If we go up a level, this is one of the problems with the whole discussion of autonomy or self-driving cars. If I pick your random person off the street and say, “What do you think a self-driving car is?”, what is generally boils down to is, self-driving cars would be absolutely great. I could get into the car, curl up in the back seat, tell it I want to go to xyz address, and then it would wake me when we’re there. And I said, “Okay, under what circumstances? Just on highways or in cities or not in snow or whatever?” “Oh! Well, no. I want it to take me wherever I want to go. Wherever I want to go, whenever I want to, in whatever circumstances. And I especially like the idea of a self-driving car if it’s snowing or blizzarding out.” Which is, of course, the hardest use case.

BRIAN SANTO: Right.

JOHN VOELCKER: So when you say, “self-driving cars,”… and this is really the argument with Tesla having named its early actively lane control product as “auto pilot,” people had the vision of something called auto pilot, which means essentially the plane is flying itself. Yeah, the pilot’s there, and if an alarm goes off, the pilot interacts. But that’s the vision for self-driving cars. And we are not there, and we are not going to be there for quite a while. Probably North America will not be there first.

BRIAN SANTO: Have you seen that video of the young man driving a Tesla with his mom in the driver’s seat and he hands her the phone and it’s like, “Here! Check this out! This is great!” And he literally climbs out of the driver’s seat into the back seat and starts pretend sleeping with a pillow.

JOHN VOELCKER: Yes. I have seen that video.

BRIAN SANTO: Nothing wrong there, right?

JOHN VOELCKER: It is sort of shorthand among some of my colleagues to refer to the two customers who have been decapitated in their Teslas. There are those who would suggest that they should have been paying attention, and hence it is their fault. There are others who say that any system that claims to be self-driving or auto pilot or what have you, that doesn’t make the driver prove that he or she is paying attention every eight to 30 seconds.

I actually test how long I can have my hands off the wheel in a regular one without getting beeped at. Some will let you go as long as 30 seconds, but that’s about the limit. Any car that doesn’t make you prove you’re at least somewhat attentive should perhaps not be sold. But I also have learned, through many scars, not to engage directly with Tesla fan boys. Life is too short.

BRIAN SANTO: Yeah. I get it.

Somewhere in the middle of our conversation, you were talking about the definitions of “autonomy.” And you said we’ll get to that later. Did we cover that in our conversation? Or did you want to return to that?

JOHN VOELCKER: We did cover it. It was essentially me saying that when you talk to the person on the street, their idea of a self-driving car is Level 5 at the highest part of the level. Anywhere, anytime, any weather. That is not how it is going to roll out.

I started at an interesting time. In 2009, the only electric car on sale was the Tesla Roadster. Barely. I ran Green Car Reports for nine years, starting then. The search terms at the beginning were “hybrid” and “Prius.” And by the end, when I left in 2018, the search terms were “electric car” and “Tesla.” If you had said to me and the rest of the auto media (absent a few sane souls who knew how to read technology) that electric cars would roll out far faster than any meaningful level of SAE autonomy, they would have smiled sadly and said, “Oh, yes. I’ve met old people like you.”

The smart people I know… I should say here, possibly the most fascinating automotive event I have ever been to and will ever be to was the DARPA Urban Challenge in 2008, which I covered for IEEE Spectrum. There are so many things that were atmospheric about it, including taking place in a repurposed Air Force base within the neighborhood that used to be the officer housing with tumbleweeds blowing across the yards and broken windows in little stucco officers’ cabins. And four layers of Jersey barriers, so that if one of the autonomous vehicles headed straight for the journalists, the Pentagon would not wipe out a bunch of journalists.

BRIAN SANTO: No! That’s more concern than I’m normally used to, but great!

JOHN VOELCKER: It was a high-profile event. It’s the kind of place where you run into GM’s head of R&D in the parking lot and you have a passionate, 45-minute chat about autonomy.

Then you could see both the promise and the immense, immense hurdles. I said that probably North America will not be on the forefront of outright autonomy simply because we have tour players here.

I don’t quote Elon Musk often, but he asked the right question many, many years ago. Against what are we judging the performance of automated safety aids in cars? If we are judging them against humans (the cyberpunk folks would say, “If we’re judging them against me”), then it is already clear that they reduce crashes, injuries and deaths. That’s why we have seen such a fast rollout, because some in the early days were so convincing between the same car, let’s say with automatic emergency braking and without, that the crash rates were lower. And that’s good right there.

But if we judge them against humans, it’s clear that automated safety systems work. If we judge them against perfection, they will not work. And I’m amazed how many people, randomly, in conversations like this, including at cocktail parties, bring up the trolley car problem. “You’re at the helm of a runaway trolley car. Do you kill one young person or three old people?”

BRIAN SANTO: Yeah.

JOHN VOELCKER: I think the answer to that is, if we ever get to the point where the car is making decisions of that subtlety, we’ve already won.

BRIAN SANTO: Here’s another question. Here’s a wrinkle. Have you heard this one? I’m sure you have. You will never eliminate traffic injury and death.

JOHN VOELCKER: Correct.

BRIAN SANTO: So you introduce cars with some level of self-control or even full autonomy. And that’s wonderful; that’s excellent and to be desired. But what you’re left with is killer robots.

JOHN VOELCKER: I must admit, I had not thought of it in that frame.

BRIAN SANTO: I’ve had these conversations with people, too. They’re like, “What happens when an autonomous car has a problem and hits somebody and you’re in the car and you’re watching and you would have not done that. But the car did.” So yes, you save 30,000 theoretical lives who weren’t hit…

JOHN VOELCKER: Tell that to the widow.

BRIAN SANTO: Tell that to the widow. Now you’ve got a killer robot. I was going to ask you if the car companies are aware of that, and you said you haven’t heard that. So they haven’t discussed that with you, obviously.

JOHN VOELCKER: I think they are extremely aware of it. Any large car company has a very large legal department. They are big targets flush with cash. It is why, quite frankly, the big companies are proceeding very slowly and very cautiously and doing hundreds of thousands, if not millions of miles of testing and only rolling out hands-off systems in places that have been very carefully mapped and that they have a very strong idea of what’s going on. And they get criticized for that all the time. “Oh, Tesla did it with a bunch of coders!” “Tesla wasn’t around 15 years ago. Look what they did.”

Tesla takes enormous risks, and some of their systems still have a “beta” label on them. If your cell phone has a beta operating system, the worst that can happen is, it crashes, it melts down, and maybe you lose your photos. If your car has a problem with its beta software, it can become a killer robot.

So I am all for, as I think most people in the auto industry are, very careful and cautious rollout. The problem is really the handover issue. Humans are hard wired with given sets of reaction times, and we know very well from lots of testing that if you are doing something else — whether it’s playing a video game or sending a text or whatever…

BRIAN SANTO: You might even be checking your rearview mirror.

JOHN VOELCKER: Most people do it (well, the ones who do it) can do it unconsciously. But they still have their hands on the wheel, and they’re mostly aware of what their car is doing. I’m thinking of something that takes both focus and brain concentration away. Under those circumstances, it takes up to eight seconds for a human to drop the phone, grab the wheel, refocus their eyes outwards, figure out why the car suddenly freaked out and said, “You must take control immediately!” and do the appropriate evasive or braking action. And at 75 miles an hour, that’s about five times too slow.

BRIAN SANTO Yeah.

JOHN VOELCKER: And so that’s why you have what some people have called “nannies” in systems like Super Cruise that really want to make sure that you are paying attention at every moment, or at least every eight seconds.

The final thing I want to say: What a lot of people don’t focus on in discussions like these (and I have an article sort of in preparation on this)… A lot of these systems are software, right? After a certain point, it becomes easier, cheaper and simpler just to build the sensors into every vehicle. But that allows you to turn them on or off via software.

In fall of 2020 or 2019 (I can’t remember now – COVID time), the very first folks who bought GM Super Cruise as an option on their Cadillacs for a four-figure sum (I forget what the sum was) for a letter in the mail saying, “Your complimentary three-year trial is up. We hope you have been enjoying Super Cruise, the most advanced system” yada yada yada. I can’t remember if it was $19.99 or $24.99 a month. “Please give us your credit card number.”

BRIAN SANTO: Ohhhh!

JOHN VOELCKER: We are going to see — and this is your fair warning, you and your listeners — we are going to see a whole host of advanced software-driven systems in the cars of the 2020 becomes pay-as-you-go. Whether it’s radio stations… think the cable package, right?

BRIAN SANTO: Right.

JOHN VOELCKER: Whether it’s advanced safety systems (although I think government regulators will have some fairly strong thoughts on that), but certainly a lot of the consumer convenience features up to and including, as one friend pointed out, “Well, suppose you didn’t pay the fee and none of your electric windows are self-closing? You have to hold the switch all the way up to close any of them, unlike modern cars that have mostly automatic switches. You’d pay $.99 a month for that convenience, wouldn’t you?”

BRIAN SANTO: Ah… Mobility as a service.

JOHN VOELCKER: Bingo.

BRIAN SANTO: John, thank you so much for your time. I appreciate it.

JOHN VOELCKER: Thank you for having me. It’s great to talk to you. We will catch up offline and do things we would not talk about on the podcast.

BRIAN SANTO: It’s a deal. All right, buddy. Take it easy.

JOHN VOELCKER: Thanks for having me. See ya.

BRIAN SANTO: That was John Voelcker, independent automotive writer and consultant.

Early in our conversation, JV suggested fact-checking him about his recollection that Mercedes Benz introducing the first adaptive cruise control – or ACC – with its Distronic technology. The company introduced Distronic in 1998. At the time, the company claimed Distronic was the first radar-based ACC. The company that is widely acknowledged to have introduced the first ACC of any type is Mitsubishi in 1992.

Another loose end from our conversation: JV mentioned Ford’s recent introduction of its hands-free technology, which the company said it will push out as an over-the-air update for some F-150 trucks and some Mustang Mach-E cars. Ford calls the feature BlueCruise.

According to Ford, BlueCruise will allow hands-free driving on “prequalified sections of divided highways called Hands-Free Blue Zones that make up more than 100,000 miles of North American roads.”

As John said during our conversation, the first instances of autonomous driving will occur first in very specific, well-mapped areas, and highways are the easiest places to do it.

We have a link on this podcast’s web page to a story that JV recently wrote for Car & Driver about California’s proposal to ban cars with gasoline engines by 2035.

We also have a link to our own EE Times article in which Edge Case Research co-founder Phil Koopman suggests that when the auto industry talks to consumers, it should ditch any mention about SAE levels of vehicle automation and talk instead about driver responsibilities.

And that is it for the Weekly Briefing for the week ending April 16th. Thank you for listening.

The Weekly Briefing is available on all the major podcast platforms, but if you get to us via our web site at eetimes.com/podcasts you’ll find a transcript along with links to the stories we mentioned.

This podcast is produced by EE Times. It was Engineered by Taylor Marvin and Greg McRae at Coupe Studios. The Segment Producer was Kaitie Huss.

I’m Brian Santo. See you next week.

Electronics Engineering Tools

原理图工具

射频工具

无源工具

PCB印刷电路板工具

数学计算器