Can We Admit Hands-Free Driving Is For Texting
It was three minutes into using Ford’s new BlueCruise “handsfree” technology that I reached for my phone to check my messages. I wasn’t using my hands to steer the F-150 Lightning that I’d been lent for my week in Colorado, so what was I supposed to do with them?
But, remembering I’m a professional and a good, attentive driver, I caught myself, put the phone down, and awkwardly rested my hands on my knees like I was being interrogated.
Six, maybe seven minutes later of uninterrupted BlueCruis’ing down Interstate 25 south towards Colorado Springs, I grabbed the phone again, and opened my email. I’d just gotten off a long flight and there was a lot of it. After about 20 seconds, the truck, which features a Driver Monitoring System (DMS) flashed a warning on the steering wheel reminding me to pay attention to the road, and I put the phone back down again. I then learned that if you think most driving is boring, try not driving, but also doing nothing else at all, besides being responsible for catching any mistakes, or you might die.
Beginning with Tesla’s inappropriately-named “Autopilot” feature in 2014, cars have been inching, slowly, somewhat deceptively, but surely, towards partial autonomy, under the guise of improved safety. Tens of thousands of people die on our roads every year, and so of course we want to stop that from happening, right? And who among us is going to argue for less saferoads, right?
I’ve spent the past few weeks using the newest “Level 2” Advanced Driver’s Aid Systems (ADAS). I’ve used Ford’s BlueCruise in Colorado and Los Angeles and I used Cadillac’s SuperCruise in New York and Connecticut. I’ve also spent lots of time with Tesla’s Autopilot, and researched pretty extensively the good and the bad of turning over controls of our vehicles to computers.
I’ve come away incredibly skeptical that the best way to make our roads safer is to invent a system whereby some cars will drive themselves some of the time. I am however certain that, if you can afford a luxury car, and your actual goal was to use your phone, read the news, or do some other task unrelated to driving, while driving, setting up a blanket of “safety” to prop this technology up against would be excellent cover.
Disclaimers: 1) I am not against the idea of autonomous vehicles. I think there are several use cases where they would come in handy and improve humanity. 2) My “test” is in no way scientific. I drove different vehicles on different roads, in different states in order to develop a general picture of the capability of the software, not to develop a controlled testing environment. This is my opinion, based on my experience and the written expertise of others in the space.
BlueCruise, SuperCruise, and Autopilot are all based on what is fundamentally the same technology: a combination of cameras and sensors in and around the vehicle can create enough of a picture of the road and the cars on it that the vehicle can perform certain maneuvers under certain circumstances without intervention from the driver. BlueCruise and SuperCruise are geofenced (or, geographically limited) to certain areas where they are known to be the safest: divided, single-direction highways with clearly marked lanes, and only in clear weather. General Motors has mapped out over 200,000 miles of highways, whereas Ford has a bit less, over 130,000 miles of highways, where these systems will work. GM uses commercial test vehicles to create lidar maps of these roads which are then downloaded by individual vehicles over the air, in order to create the clearest picture of the world.
Tesla does not geofence their system at all, deferring to the user where to activate the system (at their own risk, and the risk to everyone else on the road, of course). All three of these systems place the ultimate responsibility for the vehicle’s operation on the driver. Because Tesla has the fewest geographic safeguards, I actually feel the least safe using their system, and would not use it anywhere except a divided freeway. I’ve gone up to 10 minutes handsfree in a Tesla on the highway before being prompted to put my hands on the wheel. Though current model year Teslas do have DMS cameras (that Consumer Reports called “inadequate”), the slightly older model I drove had only the torque sensor on the steering wheel.
In Colorado, over the course of a week, I went for a maximum of about 22 minutes without having to intervene in the F-150 Lightning; the truck prompted me to take over when, due to a construction zone, the lanes shifted and painted lines became unclear. When I went to pick up my phone, one of the two DMS cameras understood what I was doing and prompted me to resume driving within about 20-30 seconds. For a computer to take over the task of driving for 22 minutes and not have any scary near-misses, crashes, or other hiccups is very impressive!
Unless you compare it to a human, and then it would be a total non-event and a very basic skill so simple it’s not even on the driving test.
SuperCruise, as fitted to a 2022 Escalade Sport, was even better than BlueCruise, with the winding, narrow Parkways of the New York area posing more of a challenge than a straight, flat, wide Colorado freeway. SuperCruise performed smoother inputs and maintained its position in the lane even when the road was relatively windy and undulating, rarely “bouncing off” the line markings. If prompted with a tap of the blinker stalk, the system would seek an opening and (conservatively) change lanes. For over 20 miles on the relatively straight, well-striped I-684 in Northern Westchester, I drove my mother to lunch using SuperCruise the whole way, and she didn’t even notice the system was on.
However, the DMS seems to be calibrated far differently than BlueCruise. In an attempt to set off the DMS to trigger a disconnection, I pretended to play with my phone in a very conspicuous way. Even after 60 seconds, SuperCruise did not disengage.
While the opportunity didn’t arise for BlueCruise, SuperCruise’s weakness seems to be relatively tight radius corners on parkways. I had three disengagements in virtually identical circumstances, right at the middle of tight parkway corners - one on the Hutchinson River Parkway, one on the Merritt Parkway, and one on the Southern State Parkway.
To its credit, Tesla’s Autopilot seems to be better at navigating these types of corners on highways, although I did try to activate it on a windy canyon road once. The system did allow me to engage, though it almost crashed immediately disengaging at the entrance to the first tight bend.
The best part of both BlueCruise and SuperCruise are that they both know with reasonable certainty where they will not work, and they tell you so. Neither system will work on undivided roads, or roads with traffic lights. They don’t make navigational decisions or follow a programmed navigation route. If they approach an area with uncertain map data or unclear road markings, they will prompt a warning to take over. I experienced this on I-95 heading north at the New York / Connecticut border. A recently completed 10-year project renovating a 2-mile corridor had not yet been mapped, and the Escalade told me as such. One I went back to the non-renovated section north of the border, it prompted re-engagement.
Both of these systems tend to err on the side of conservative, which, in my opinion, is good. It does not help the future of AV’s to oversell the ability of these systems, to make people think the cars drive themselves, or to imply they are safe to use in all circumstances and all locations. Tesla has been overselling the capability of their cars for years, making for an excellent opportunity for people to a make money off this lie in the stock market. Overconfidence in their Autopilot system has directly or indirectly lead to many reported crashes, and several deaths.
GM and Ford have both done admirable jobs integrating handsfree systems in a way that, at least from the drivers seat over a couple hundred miles of travel, seem reasonably safe. But from both a legal and a moral perspective, you, the driver, are still expected to be attentive and ready to take over at any time. Your hands are free…. to do nothing. Your eyes are free … to still watch the road.
Which begs the question: what is the point, exactly? There is no evidence that handsfree systems are safer than just driving the car. If you are literally unable to keep a modern car in a lane on a freeway, I’m sorry, but you have no business driving a car at all. This is a very simple task that an undistracted human is perfectly adept at.
Proponents of this tech like to lean on the safety argument, comparing the safety of L2 ADAS to some “average driver,” as if a human driver couldn’t somehow improve their own abilities to be safer behind the wheel. I have yet to see anyone compare this tech to a hypothetical exceptional driver who has driven for 40 or 50 years without so much as a moving violation or accident. They never say what I really think is the truth: that it makes it safer for you to play with your phone on the highway.
I am among the hardest core car enthusiasts that exist on planet earth, and I can admit the average commute on the average day is not the kind of driving I look forward to. Roads are in worse condition and more congested than in recent memory, and our phone addictions are rapidly bleeding into our commuting time, with our cars mirroring our phones in many ways.
And there’s the speed that work-related communication happens today, in which millennials and GenZ are accustomed to being in contact virtually 24/7. It is really really hard to have to drive somewhere on a busy work day and pretend like your phone isn’t there. We all are guilty of using our phones at some point while driving. Or we’re eating, putting on makeup, reading, or combining the act of driving with doing something you almost certainly shouldn’t be doing while driving. I’ve been guilty of it myself, though admittedly, technology like CarPlay and voice-to-text has made me communicate in a safer manner than actually typing on a small touch screen.
If there is a technology that allows you to become a true passenger, in which not only can you doom scroll Twitter, read a book, or take a nap on those long commutes, but also the manufacturer of the system accepts liability once it’s been activated, I’m all in. If there is a technology that allows me to take my own car to the bar, polish off a bottle of bourbon, press a button, and get home safely and legally in that same car, I’m all in. That sounds nice, and it would improve my life by allowing me extra time in my day to do something other than burn precious hours at mundane driving tasks.
Last month, the IIHS came out with a study in which they polled 1,010 drivers about their feelings related to partial autonomy and handsfree capabilities. In the results, eighty percent of drivers want their vehicles to have some kind of lane centering, and 27% want to be able to drive at least some of the time handsfree. Of those, 46% say that they prefer handsfree even though it will make them “much more likely to do non-driving related activities.” 32% said “somewhat more likely,” and 14% said “just as likely. So of the people who prefer handsfree driving, 92% of those people are likely to do non-driving tasks, or misuse this technology as it is designed. Even though they are aware of the dangers of doing so.
Let’s be real: Would this tech be safer than me simply being attentive while commuting? Would it be safer than me taking a cab to the bar and then home? That remains to be seen, but odds are, no, not really. What about safer than having more rigorous vehicle inspections, improving urban road design, or variable speed limits based on conditions? Would it be safer than improving our driving standards, having mandatory re-testing, or any other number of already-proven methods of reducing road deaths that don’t involve inventing whole new technologies? That remains to be seen. But we are, it seems committed to computers being better than the “average driver,” while still very imperfect, and without being willing to improve what the term “average driver” means, heaven forbid looking at road design as a contributing factor.
So let’s just call it like it is: handsfree driving is a toy that makes texting and driving on the freeway somewhat safer than texting and steering with your knees, which is what you were doing before.
You Might Also Like
