Transportation

How Misperception Stokes Taking An Offramp At Dangerous Speeds, Posing Dire Consequences For Human Drivers And AI Self-Driving Cars


You may not know this word, but you most certainly know its meaning.

Velocitation.

What’s that, you might be wondering?

It is the somewhat curious and altogether menacing phenomenon of misperceiving your speed, particularly encountered when driving a car and doing so at relatively high speeds.

Here’s the deal.

You are minding your own business and driving across town via the everyday freeway system. Luckily, there isn’t much traffic. You are zipping along at the allowed posted speed of 65 miles per hour, and admittedly juicing things a bit by going perhaps five to ten miles per hour faster. All in all, you are merely keeping up with the prevailing traffic.

In fact, to your credit, there are occasional cars that pass you, which means those drivers are going maybe 80 to 85 miles per hour. That seems excessive. Of course, it is downright illegal too. In any case, you watch them as they quickly come up from behind you and then zoom past you.

It is almost as though you are standing still or perhaps moving at a turtle’s pace.

When you are in midst of traffic which is cruising along at your speed and on occasion somewhat faster, the whole ballgame is about relative speeds. You forget that all of you are moving at a high rate of speed. There is a semblance of relative speed versus thinking about the absolute speed that comes to the fore on this matter.

In those situations of everyone moving fast, your focus tends to be preoccupied with the relative differences in speeds.

To get a sense of this difference, imagine that you were standing on the side of the freeway and watching the traffic go by. A person standing still would perceive the cars as going monstrously fast. Each car would seemingly move like the wind and gush right past the unmoving observer. Those cars are all racing at breakneck speeds, so it seems.

On the other hand, when sitting in the driver’s seat, and if you are moving at the same speed on the freeway as nearby cars, you don’t get a visceral sense of going fast. You are all moving at the same general pace. Thus, when a car going maybe ten to twenty miles an hour goes past you, it almost seems as though you are standing still. The reality is a lot different and you are indeed proceeding at a breathtaking clip.

For my discussion in detail about perception and relative speeds, see the link here. There is an Einstein-like facet of some decidedly delightful and insightful fascination involved.

In any case, if you were to stay on the freeway for any substantive amount of time, you would gradually and ostensibly inexorably get used to the prevailing speed. This perception can be rudely altered or disrupted by for example getting stuck in crawling traffic that might cause you to dramatically reduce your speeds. Those bumper-to-bumper moments are bound to force you into a series of slow and fast measures, alternating as the traffic situation allows.

That knocks you out of the zoned mindless sense of a constantly high speed.

But when the speeds are steady, unbreaking, and proceeding at a top-notch pace, your mind begins to get settled into the rhythm of the speed. It is almost as though you undergo a subtle and subconscious mental calibration resetting your norms to the fast tempo. You are in the zone, as it were. To some degree, you might not even be fully cognizant of how fast you are going.

Have you ever glanced down at your speedometer and been somewhat surprised at your speed?

That can readily happen when you are in the mental state of becoming numb to your existent pace. You were going at 55 miles per hour. Then, over a few minutes gradually increased to 65 mph. And then, you find yourself now going 75 mph. If this was done smoothly and not in any jerky efforts of acceleration, you might be shocked to notice that you are doing 75 mph when you felt like it was still 55 mph.

For getting into that mental zone, there usually can’t be much if any acceleration taking place per se. Any rapid or jumpy acceleration can potentially snap you out of the fluid mindset. The zone seems to most often occur when you are simply maintaining a nearly constant fast speed. For my analysis of what happens to a driver’s perceptions when acceleration enters into a driving journey, known as accelarousal, see the link here.

Okay, we’ve got the notion established that you can become complacent about your speed and feel as though you are not moving as fast as you really are. You might perceive your speed to be slower, possibly a lot slower, than the actual speed underway.

So what?

Well, this is where the dynamics and dangers of velocitation come to play.

Imagine that you’ve been driving like this for quite a while. Getting across town takes about 45 minutes or so of driving time. During the entire driving journey, you have had the good fortune of wide-open traffic and a constant full-throttle speed of about 70 miles per hour.

Easy-peasy.

Up ahead is the freeway exit that you want to take.

The posted speed limit for the offramp is 45 mph. This means that you are supposed to make sure that when you are engaged into the offramp that you try to be at 45 miles per hour or less (as suitable to existing roadway conditions). For example, if it is raining and the roads are slick, you would likely want to be going much slower, perhaps 35 mph or whatever makes sense for the situation at hand.

Let’s pretend that you were going 55 mph on the freeway and began to take the offramp. At that juncture, you are doing about ten miles per hour faster than the posted speed. That’s probably okay, as long as you are in the midst of reducing your speed to the indicated 45 mph. We are assuming in this scenario that it is a sunny day, the roads are nice and dry, etc.

Even if you remain at 55 mph for a tad long, you are probably still doing satisfactorily. You might find yourself getting a bit overstretched as you near the end of the offramp. Of course, if the offramp has a sharp curve to it, you will undoubtedly feel the heightened speed by the pulling forces of physics that will be trying to take the vehicle afield of the offramp. You don’t want that to happen.

All in all, being around 5 to 10 miles per hour over the posted speed of the offramp might be livable, though certainly not recommended and you are perilously asking for potential troubles.

Let’s recast the scenario and have you going at say 70 miles per hour.

I think you know what this is going to produce. If you take a 45 mph offramp at 70 mph, and for which you are now doing a strident 25 miles per hour excess speed, the world is not going to take this kindly. The odds are that you’ll end up doing a massive heart-stopping braking action when you reach the end of the offramp. You might skid into the intersection or streets that are abutting the offramp. This is a car crash in the making.

Equally worse, when the offramp has a curved shape, you are playing with fire in that the car can go flying off the offramp. That might sound exciting, but it is not. It is a path to severe injury and potential death. You are encouraged to look online at the numerous news reports about drivers that misjudged a curved offramp and the deadly catastrophe that ensued.

You might be tempted to say that nobody in their right mind would end up taking an offramp at such an abundant speed. Only those that are drunk drivers would do so. Possibly a drowsy driver might also make that blunder. A distracted driver that is watching cat videos on their smartphone could potentially find themselves doing the same mistake.

Certainly not any rational driver that has their mind devoted to the driving task and that is diligently paying attention to the roadway.

Regrettably, I dare say, velocitation can take a chunk out of that assertion that a rational driver of a diligent nature would never misjudge an offramp in the manner heretofore depicted.

Your mind can sometimes fool you, as you likely know. Have you ever seen someone in a crowd that looked like an old friend that you haven’t seen for years? You are absolutely sure it is the same person. Turns out that when you get closer, you realize that it is not the same person. An active and aware mind can potentially take what it sees and mishmash it into something else.

With velocitation, you are so accustomed to the fast speed that has become routine during a driving journey that you no longer adequately can judge your actual speed. The 70 mph feels like 55 mph, or maybe even 45 mph. Your eyes are on the roadway and the offramp, and not looking at the speedometer. You feel like there is no need to look at the speedometer since you can feel the existing speed in your bones (so you falsely believe).

After you get immersed in the offramp, you are now going to begin to realize that there is something askew about your speed and the confines of the offramp. Oops, you realize, I’m going much too fast for this offramp. Time to bleed down some speed.

Most of the time, you probably can squeak through and get enough speed off the vehicle to still complete the offramp without a devastating result. It can be a moment though of sheer terror. Assuming you safely finish the offramp, you are sweating profusely and probably swear an oath that you will never let yourself be taken in again.

The thing is, velocitation is somewhat insidious.

Some people don’t know that such a phenomenon exists. They are oblivious to the possibility of it. The instances of taking an offramp too fast are usually chalked up as a kind of oddity or quirky circumstance. Those drivers that are unfamiliar with this mental perception issue are assuming that any such occurrence is a fluke, a one-time occasion.

Other people know about the aspects of velocitation, perhaps vaguely so, and might not realize there is a word devoted to this particular manner of concern. Yet, despite the awareness of the potential for misjudging their speed, they make the same error repeatedly.

That’s oftentimes because drivers tend to overrate their own driving prowess.

Sure, maybe you got caught by velocitation a month ago, but you are better than that and know that this won’t grip you ever again. Lo and behold, a trip across town happens and perhaps the same thing occurs a second time.

What can help to break out of that mental zone would be any foul memories about that particular offramp. Thus, if you were to take that same offramp at a later date, your mind might flashback to the terror of previously using that offramp. In that case, you might slow down adequately. In fact, you might overcompensate and essentially take your car to a crawling speed on that exit.

Your mind might not generalize the matter to other offramps. This suggests that if you drove across town and decide to take a different offramp, you could get tricked again. You have flagged in your mind the dangers of the other offramp. This newly being used offramp is clean in terms of your memories and not bound to spark any concerns.

Another factor that comes to play is the nature of your car.

In the earlier days of automobiles, the engine noise would be a key differentiator about your prevailing speed. While driving a car, you could pretty much gauge the speed by listening to the sound of the engine. No need to look at the speedometer. Your hearing was enough to tell you all that you needed to know about your pace.

Likewise, cars might rattle or prattle as your speed got faster. This added an additional sound as a clue to your speed. Furthermore, you could feel the speed rattling your entire body. When your driver’s seat was shifting back and forth or shaking, that was a sign of your speed. Your body and limbs could feel the essence of the speed.

Modern cars are usually built so solidly that few if any of those cues are still identifiable. You probably won’t hear the sound of the engine as you get to higher and higher speeds. There is plenty of specially designed soundproofing materials that deaden the sounds from coming into the interior of the vehicle. The shock absorbers and other contraptions are provided to ensure a smooth ride.

All in all, you would be hard-pressed to ascertain your speed via something that the car is doing, other than by looking at the speedometer.

As mentioned earlier, you would have other important clues as evidenced by the other traffic around you. When nearby traffic is going at a widely varying array of speeds, you are probably going to be a lot more cognizant of your speed. If there isn’t any other traffic, you can be embraced by velocitation, and similarly if there is constant and seamless traffic the same can occur.

I trust that you are willing to agree that velocitation exists and that drivers need to be mindful about getting sucked into the void of misjudging their vehicle and its speed.

Let’s next consult our crystal ball and gaze into the future.

Consider that the future of cars consists of AI-based true self-driving cars.

There isn’t a human driver involved in a true self-driving car. Keep in mind that true self-driving cars are driven via an AI driving system. There isn’t a need for a human driver at the wheel, and nor is there a provision for a human to drive the vehicle. For my extensive and ongoing coverage of Autonomous Vehicles (AVs) and especially self-driving cars, see the link here.

Here’s an intriguing question that is worth pondering: Is it conceivable that AI-based true self-driving cars might be susceptible to velocitation, and if so, what can be done about this?

I’d like to first further clarify what is meant when I refer to true self-driving cars.

Understanding The Levels Of Self-Driving Cars

As a clarification, true self-driving cars are ones that the AI drives the car entirely on its own and there isn’t any human assistance during the driving task.

These driverless vehicles are considered Level 4 and Level 5 (see my explanation at this link here), while a car that requires a human driver to co-share the driving effort is usually considered at Level 2 or Level 3. The cars that co-share the driving task are described as being semi-autonomous, and typically contain a variety of automated add-on’s that are referred to as ADAS (Advanced Driver-Assistance Systems).

There is not yet a true self-driving car at Level 5, which we don’t yet even know if this will be possible to achieve, and nor how long it will take to get there.

Meanwhile, the Level 4 efforts are gradually trying to get some traction by undergoing very narrow and selective public roadway trials, though there is controversy over whether this testing should be allowed per se (we are all life-or-death guinea pigs in an experiment taking place on our highways and byways, some contend, see my coverage at this link here).

Since semi-autonomous cars require a human driver, the adoption of those types of cars won’t be markedly different than driving conventional vehicles, so there’s not much new per se to cover about them on this topic (though, as you’ll see in a moment, the points next made are generally applicable).

For semi-autonomous cars, it is important that the public needs to be forewarned about a disturbing aspect that’s been arising lately, namely that despite those human drivers that keep posting videos of themselves falling asleep at the wheel of a Level 2 or Level 3 car, we all need to avoid being misled into believing that the driver can take away their attention from the driving task while driving a semi-autonomous car.

You are the responsible party for the driving actions of the vehicle, regardless of how much automation might be tossed into a Level 2 or Level 3.

Self-Driving Cars And Velocitation

For Level 4 and Level 5 true self-driving vehicles, there won’t be a human driver involved in the driving task.

All occupants will be passengers.

The AI is doing the driving.

One aspect to immediately discuss entails the fact that the AI involved in today’s AI driving systems is not sentient. In other words, the AI is altogether a collective of computer-based programming and algorithms, and most assuredly not able to reason in the same manner that humans can.

Why is this added emphasis about the AI not being sentient?

Because I want to underscore that when discussing the role of the AI driving system, I am not ascribing human qualities to the AI. Please be aware that there is an ongoing and dangerous tendency these days to anthropomorphize AI. In essence, people are assigning human-like sentience to today’s AI, despite the undeniable and inarguable fact that no such AI exists as yet.

With that clarification, you can envision that the AI driving system won’t natively somehow “know” about the facets of driving. Driving and all that it entails will need to be programmed as part of the hardware and software of the self-driving car.

Let’s dive into the myriad of aspects that come to play on this topic.

The first place to start this discussion consists of clearly acknowledging that today’s AI driving systems are not sentient and therefore cannot be subject to velocitation in the same manner as human beings. In short, if you strictly assert that velocitation requires human eyes and a human brain, we can summarily then reject the possibility of self-driving cars as being susceptible to velocitation.

Case seemingly closed.

But wait for a second, perhaps we can use the velocitation conceptually to glean some insights into what might go awry with self-driving cars.

Is it possible that a self-driving car could take an offramp at too high a speed?

Yes, absolutely so.

Let’s explore the various ways that this could happen.

One obvious aspect would be that the AI driving system could potentially fail to have an accurate indication of what the speed for a given offramp is supposed to be. If that were the case, it is conceivable that the self-driving car could be moving along at the prevailing freeway speed as it comes onto and uses the offramp. The AI driving system would be awaiting some other indication that the speed should be lessened, otherwise the existing speed prevails.

I’m sure you are already clamoring that the AI driving system ought to have digital maps that include the offramp and therefore would have a stated speed that is associated with the offramp. As such, the odds are that the AI driving system would be able to readily direct the driving controls to reduce the freeway speed of the self-driving car accordingly as it utilizes the offramp.

Though most automakers and self-driving tech firms are indeed making use of extensive digital maps, this is not always necessarily the case. In addition, there is a chance (admittedly slim) that the digital map has an inaccurate indication about the offramp speed.

Okay, you might say, but certainly, the self-driving car is going to detect a roadway sign that showcases the allowed speed for that offramp.

Yes, it would normally be the case that the sensors of the self-driving car would presumably detect such a roadway sign. This would principally be done via the use of video cameras. Upon the video camera images being examined via Machine Learning (ML) and Deep Learning (DL) technologies, those computational pattern matching algorithms are apt to discern that there is a posted sign and also determine that the sign provides an indicated speed.

In the future, the expectation is that much of the roadway infrastructure will be making use of V2I (vehicle-to-infrastructure) electronic communications. There will be various devices such as IoT (Internet of Things) and edge computing units all over our highways and byways, see my elaborated discussion at this link here. Traffic signals will be beaming out electronic signals that self-driving cars can receive and ergo be alerted to whether the light is red, green, or yellow. Bridges can send out V2I that indicates when the bridge is closed or otherwise unavailable. Etc.

You can anticipate that roadway signs such as posted speed limits will also be part of the V2I efforts. Those signs will be beaming out an electronic signal indicating the speed in that area. In addition, the speed could be variable so that at some times of the day it has a “posted” speed of say 35, while at other times of the day a different speed is stated.

Anyway, that’s more about the future and not what we have in hand today.

We were discussing that the self-driving car ought to be using its sensors to visually discern the offramp stated speed via capturing images of the driving scene and looking for an appropriate posted sign. This would seem to dispense with any reliance on a digital map since the posted sign would seemingly be the final arbiter of what the speed is.

The problem with relying solely on a posted sign is that the sign might be rigged or it might not be there. In the case of not being there, suppose the posted sign got struck by a wayward car and got knocked down to the ground, and has not yet been reposted. That means there isn’t any posted sign to be readily discovered (it is laying in the dirt and marginally visible).

In terms of being rigged, there are ongoing concerns that self-driving cars can be potentially tricked in rather simple and highly dangerous ways. By masking a stop sign or altering the numbers on a speed limit sign, it is feasible to fool the AI driving system into making mistakes about the driving conditions. A human driver uses common sense (most of the time) to realize when something is amiss. Efforts to imbue AI systems with common sense are ongoing but have been sorely lacking so far (see my analysis at this link here).

All in all, the emphasis is that a self-driving car might not have a proper indication about the offramp permitted speed. It might even have a false indication that was purposely crafted by evildoers to beguile self-driving cars. Generally, this is seemingly unlikely at this juncture and we can assess that it won’t happen with any frequency.

Nonetheless, it can happen and sadly will happen.

Another possible way to have a self-driving car falter when using an offramp would be to have the AI driving system fail to ascertain that the offramp is an offramp.

How so?

Consider something that maybe has happened to you.

You might have been using your GPS and mapping software to drive in an area that has lots of freeway lanes and offramps and noticed that sometimes the system gets somewhat confounded. For example, a lane that branches off the freeway onto a different freeway can be the target lane that you are supposed to take. Imagine that you inadvertently don’t make it into that lane. You’ve perhaps seen instances when the GPS and mapping system portrayed your location as though you were on that branching lane, even though you were still on the regular lane of the freeway.

In the case of a self-driving car, suppose the GPS and mapping were showcasing the autonomous vehicle as though it was still on the freeway, despite the fact that the AI driving system is maneuvering the self-driving car onto the offramp. This momentarily can create some difficulty for the AI driving system. Presumably, the sensors are indicating that the self-driving car is no longer on the freeway per se, meanwhile, the map is indicating to the contrary.

This could lead to the potential of the self-driving car not reducing speed as matching to the needs of the offramp.

Here’s yet another possibility.

Some self-driving cars are programmed to do what insiders refer to as a follow-the-leader driving approach. The AI driving system detects a car ahead of it. If the car ahead slows down, the AI driving system will slow down the self-driving car. If the car ahead speeds up, the AI driving system will speed up the self-driving car.

Envision a human-driven car that opts to take the offramp and the human driver is experiencing a bit of velocitation and going too fast. A self-driving car is immediately behind the human-driven car and also taking the same offramp.

At that juncture, the AI driving system might calculate that it is okay to match the speed of the car ahead. Whether the AI driving system also has detected the posted speed sign or maybe has digital maps, could get the algorithm to computationally ascertain that following the human driver in this manner is inappropriate. Thus, the follow-the-leader is momentarily overridden.

On the other hand, it is also possible that the follow-the-leader is the “best” data that the AI driving system has. In that case, like lemmings going off of a cliff together, there is a danger that the self-driving car will follow the human-driven car into potential doom.

Numerous other possibilities could get the self-driving car into trouble about the offramp.

Suppose the self-driving car has some kind of bug or error in the software. This can happen. Keep in mind that there are going to be perhaps millions upon millions of lines of code, ostensibly hundreds of millions, and there are lurking errors that have yet to show themselves.

A bad bug at the wrong time could get the self-driving car into an out of sorts status as it proceeds down the offramp.

Conclusion

I’ve somewhat focused on velocitation when it comes to using a freeway and taking an offramp, but that’s just a particular use case (a quite common one).

Velocitation can arise in other ways.

You might be on a lengthy highway. You might be on a long stretch of streets. Those two have the same potential issues involving getting accustomed to a speed and misjudging what your actual speed is. If you are at some juncture wanting to make an upcoming left or right turn, your realization of your present speed versus the speed required for the turn can be askew.

For humans, velocitation can readily arise and catch us quite off-guard. Surprisingly, those that know about this vexing mental perception vulnerability can still fall into its dangerous and potentially deadly trap. That is the very definition of being insidious, it would seem.

AI self-driving cars can have similar kinds of difficulties, which we would not want to especially label as velocitation, due to the desire to avoid anthropomorphizing the AI. That being said, it is useful to consider the myriad of human foibles at driving and see what we can learn from how humans make mistakes while at the wheel.

Those human guffaws are surely helpful for trying to make AI driving systems be the best drivers that they can be.



READ NEWS SOURCE

Leave a Reply

This website uses cookies. By continuing to use this site, you accept our use of cookies.