New Products For Pedestrian “See Me” Wearable Safety Gear Being Spurred By Self-Driving Cars

Pedestrians have a tough life.

When trying to cross the street, you ostensibly are taking your own life into your hands as you dare to go in front of menacing car traffic. Even if a slew of cars is standing still at a red light while you lawfully use the crosswalk, there is still a chance that a driver might mistakenly use their gas pedal and suddenly accelerate forward into you.

Always keep your eyes on those momentarily stationary cars and their drivers, just in case you can miraculously discern that a sudden movement is on the verge of occurring and potentially leap out of the way.

The scariest situations entail a crosswalk that has no traffic signals at all. These uncontrolled wildcard circumstances are a disaster lying in wait.

Will the upcoming cars come to a halt?

Do they see me?

Once they come to a halt, will they remain still until I’ve completed getting across the street?

We teach children to look both ways. This includes the sage advice to only cross where it is safe to do so. Some youngsters seem to think that if there is a crosswalk in existence, it must axiomatically be a safe place to cross the street. Perhaps these kids assume that those faded painted lines on the street can form a veritable protective barrier that is impervious to being breached by moving cars.

Of course, not all pedestrians necessarily cross the street in a legal and nor prudent manner.

Jaywalkers make a quite dicey decision by crossing at any juncture of a street, doing so at their own peril and upon the belief that they know best about where to properly cross. Sometimes a jaywalker fails to notice oncoming traffic and gets themselves into a dire posture. In other cases, the jaywalker abundantly observes the traffic and yet relies upon their adroit dancing and twisting skills to try and do a Frogger-like escape. Sadly, whichever method a jaywalker opts to use, there is a hefty chance that they will lose this game.

And losing the game of being a pedestrian is very costly.

It seems strange that there are pedestrians that seem to think they will win in a contest with a car and its car driver. The car driver has the multi-ton vehicle at their disposal and can unquestionably be the winner if you are comparing a flimsy human body to the impact destructive forces of a car in motion.

There is no debate that the human will lose, and the car will win.

I’m not suggesting that the car is in the right. The point simply is that if you weigh the physics and forces involved, and merely consider what will happen when a human is struck by a car, the car is going to apply tremendously adverse forces to the pedestrian. Unless you believe in the comic books and superpowers, the odds of a human holding out their hand and bringing the car to a shattering stop, while the human remains standing and happy, well, keep dreaming.

Yet there are daily and plenty of pedestrians that dare cars to come and get them. A pedestrian steps into the street and eyeballs a car driver. The car driver has to decide whether to slow down and possibly stop to let the pedestrian cross. Similar to two gunslingers, they get into a staring contest.

They are unequal gunslingers in terms of how they are armed.

The pedestrian figures that they are nimble and can presumably escape death by pirouetting if the situation so dictates. Perhaps the greater tool that the pedestrian has at their fingertips is the notion that the car driver will be in a lot of trouble, assuming that the car proceeds ahead and rams into the pedestrian.

This is bad news for the driver. The driver will have to deal with repairing the damages done to their car. The driver will undoubtedly have their car insurance rates rise. The driver will possibly have their driver’s license suspended or revoked. And the driver could face criminal charges.

The thing is all those special factors that arm the pedestrian seem conceptual and not especially tangible at the time of this life-or-death stare down. The reality there on the ground, in the middle of the street, says that when a human and a car collide, the human is going to be the one suffering the most hurt. Sure, you can argue that the car driver is going to possibly get into deep trouble, but that’s long after the split-second instance when the pedestrian loses by getting injured or destroyed.

The outcomes for the pedestrian are to either by the skin-of-their-teeth make it across the street untouched, or possibly get rammed and severely damaged or killed.

The preferred approach to these eye-to-eye contests can differ by locale. Some contend that in New York City, a pedestrian must try to make eye contact with the driver since otherwise, the pedestrian is basically signaling that the driver has the right-of-way and the pedestrian is a goner. Meanwhile, some advise that in Boston the best bet is to purposely not make eye contact with the driver. If a pedestrian does make eye contact, this implies the pedestrian has now lost the street-crossing gambit and the driver is cleared to proceed unabated.

Or maybe it is the other way round.

These norms about being a pedestrian are not very comforting. Suppose the driver is from out-of-town and does not know the expected convention? Another possibility is that a local and well-seasoned driver has decided to buck convention, on that particular day and time, thus they are decidedly not going to play “by the customary rules” known as sacrosanct by pedestrians in that area.

Many towns and cities have been taking proactive measures to try and stem the seemingly rising tide of collisions involving cars and pedestrians. These municipal campaigns are often referred to as Vision Zero or a similar catchy naming. The idea is that by alerting everyone to the pedestrian considerations, hopefully, the number of car collision fatalities involving pedestrians will fall to zero.

It is easy to see that there are two components to the problem.

There are the pedestrians and their desire to cross the street, and there are the cars that are using the street to proceed and wish to do so without disruption or delay. When the two meet there is a high potential for fireworks.

You can try to get pedestrians to be more observant and tread carefully when crossing the street. This is handy but it is not foolproof. People might not retain the thoughtfully urged warnings that have been plastered on billboards and social media. Posting signs at streets is certainly useful, though once again the problem entails having the pedestrian realize the sign is there, having them read the sign, and then in the throes of wanting to cross the street they need to heed the sign.

Don’t bet all your chips on that happening.

For car drivers, they are maybe going to see the myriad of signs and billboards about pedestrians and take those warnings to heart. On the other hand, when it comes down to rushing on your way to work or trying to get home to see your children, the reality is that many drivers are not especially tuned to the needs of pedestrians. We already all know that car traffic itself is a dog-eat-dog world, and the feeling extends to the ongoing warfare between cars and pedestrians.

We seem to have landed at a loggerhead of not finding a solution to this problem. Despite extensive attempts to inform and spur pedestrians to be safer, and despite the same being done with car drivers, the sad truth is that the problem fundamentally remains as is. To clarify, we still need to try and do that kind of outreach and it does make a difference, clearly so, but the point here is that the chances of reaching the vaunted zero are a longshot, at best.

What other solutions might be possible, you are likely wondering.

One part of the problem involves making sure that the car driver realizes there is a pedestrian present at the moment in time when the pedestrian and the car are about to intersect in the so-called time-space continuum, as it were.

Pedestrians can be hard to see.

They might be wearing clothes that tend to blend in with the prevailing driving scene. In terms of children, they are smaller in stature and less likely to be readily seen. At nighttime, a pedestrian can be cloaked by darkness. There are a zillion ways that a car driver can fail to spot a pedestrian.

In theory, if a car driver knows in-advance that a pedestrian is there, this would at least reduce the incidents that involve a setting whereby the driver was caught unawares and therefore did not react in time to avoid striking the pedestrian. This is not a cure-all since a driver can opt to not avoid a pedestrian, intentionally striking the pedestrian and knowing full well they are about to do so. That’s another slice of the problem and a variant that I’ll be covering in a later column.

In any case, there are already various methods used to try and make pedestrians standout such that car drivers will take notice.

Crossing guards wear brightly colored jackets. Some parents use the same kind of garb on their young children when sending them off to school. Another possibility is holding and waving a flag. You’ve probably seen those triangular-shaped flags that are fluorescent orange and mounted on a lengthy pole. A pedestrian holding the pole in an upright position will aid in being spotted from a greater distance.

Some pedestrians eschew those rather blatant options, especially since those seem to be uncool or otherwise draw the ire of cynics. A more measured approach involves having reflectors on your clothing or perhaps displayed on a hat or cap that you are wearing. These seem to be more acceptable as a form of grabbing the attention of car drivers.

Those are all helpful and certainly are encouraged as a means of startling car drivers out of their zombie-like trances and becoming conscious of pedestrians up ahead of their car. That being said, we can all likely agree that this still, unfortunately, does not guarantee that a pedestrian will be seen (it improves the odds but doesn’t push the probability to a complete one hundred percent).

Shifting gears, consider that the future of cars entails the advent of self-driving cars. For AI-based true self-driving cars, there will not be a human at the wheel. An AI-based driving system will be doing the driving.

This brings up the matter of whether self-driving cars will be better at detecting pedestrians than do human drivers. Some worry that self-driving cars might be worse at pedestrian detection. That would be disconcerting, for sure. We would seemingly hope and assume that self-driving cars would be at least as good as humans, and indeed be relishing the possibility that self-driving cars would do even better at the pedestrian detection task.

Human drivers are apt to be distracted while driving. The emergence of now-ubiquitous smartphones has become a key distractor while driving. Some pundits point out that human drivers watching cat videos or trying to text while driving are a likely basis for the increase in pedestrian car-related fatalities. Humans also drink and drive.

AI-based driving systems aren’t going to be drinking and driving. Nor will the AI driving systems be driving in a distracted mode akin to how humans lose their focus or opt to look away while at the wheel. In that sense of things, it seems obvious and natural to anticipate that self-driving cars will do a better job at detecting pedestrians.

If self-driving cars can do a better job at detecting pedestrians, this suggests that there will be a lesser number of pedestrian car-related fatalities and striking incidents.

Some wonder whether the wearing of brightly colored jackets and reflectors will still be needed in an era of self-driving cars.

Can we put away those contrivances and just assume that the AI driving systems will be able to properly detect us?

The short answer is no, you cannot ditch those helpful precautions. They will still be useful and recommended for safety purposes.

Some go even further, asserting that you might want to add some new and additional means to be readily spotted by self-driving cars. In other words, there are means of being spotted by human drivers, which can also apply to being spotted by self-driving cars. Also, there are ways to be spotted by self-driving cars that are specific to the nature of self-driving cars (and might or might not apply to human drivers per se).

This is today’s intriguing question: Will the market of pedestrian garb and contraptions for warning human car drivers be spurred toward new products to cope with the inclusion of true self-driving cars on our roadways?

Before we unpack the matter, let’s first clarify what I mean by referring 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 Pedestrians

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.

Detecting other cars that are nearby to a self-driving car is somewhat straightforward for the AI driving system. Usually, other cars are relatively sizable, and a self-driving car can make use of its on-board video cameras, radar, LIDAR, ultrasonic units, and other sensory devices to pretty much figure out that a car is in the driving scene of interest.

Pedestrians are another beast altogether.

Trying to detect a pedestrian can be quite tricky.

As I’ve previously discussed, some AI driving systems are for example at times categorizing a snowman as being a human pedestrian (see my analysis at this link here). The point being that pedestrians are not customarily as easy to detect as a car, and yet we are certainly going to be expecting that AI driving systems will treat pedestrians with as much care, respect, and rapt attention as they would nearby cars (maybe that is a somewhat argumentative claim, but you get the gist).

The primary means of detecting pedestrians involves the video cameras of the self-driving car.

Streaming video is being collected while the vehicle is underway. Algorithms inspect the video images and are attempting to discern whether there are any blobs in the imagery that appear to be pedestrians. A pedestrian might be standing behind a light pole or some other obstruction, thus the image might have just a partial indication that a pedestrian is present. It can be arduous for computer-based image processing to do a Where’s Waldo inspection of video to identify where a pedestrian might be in the cacophony of objects in a busy street scene.

A child might be standing behind an adult, perhaps not seen at all from the vantage point of an oncoming self-driving car and its video cameras.

Imagine that the child opts to suddenly and unexpectedly dart from behind the adult and straight out into the middle of the street. Since the AI driving system did not detect earlier that the child was present (being hidden behind the adult), this means that the appearance of the child will be detected later in the game, so to speak. This in turn might reduce the viable options of trying to come to a halt versus having to veer away from the child that is now within striking distance.

Would wearing a traditional fluorescent vest or jacket help (the kind used to alert human drivers)?

And what about carrying a bright orange flag that was mounted on a tall pole?

Yes, those are still good to go.

Assuming of course that the image processing algorithm being used takes this into account.

Anything that would visually catch the eye of a human driver is likely to be advantageous for being exploited too by self-driving cars. The key is that the automaker and self-driving car developers need to ensure that they are incorporating those telltale clues into the image analysis and pedestrian detection programs that they are using.

This includes the use of Machine Learning (ML) and Deep Learning (DL). To utilize today’s ML/DL, you need to train the computational pattern matching algorithms via lots of data. Thus, it would be important to include images of pedestrians wearing the type of “see me” garb that is being used to attract the eye of human drivers. The pattern matching by the ML/DL would hopefully usefully ascertain that those are handy cues signifying a pedestrian is present.

One difficulty with those pattern matching approaches is that they can be either purposefully fooled or inadvertently fooled.

Suppose that a fluorescent jacket is placed on a mannequin. The mannequin is placed on the sidewalk, perhaps outside a store that is selling clothing. It is conceivable that a self-driving car that is passing along that street might misinterpret the mannequin as being a human pedestrian. Why? One possibility is that the pattern matching “learned” that whenever there is a fluorescent jacket on a figure that has the overall shape of a chest and arms, this implies that there is a person there and thus categorized as a pedestrian.

Humans would be less likely to make such a categorization error.

Of course, as previously mentioned, humans make a lot of other kinds of driving errors, thus they do not get a free pass simply because they can usually tell the difference between a mannequin and a real human.

Some have suggested that maybe we ought to come up with a new symbol that can be standardized as a universal “I’m a pedestrian” symbol and then make sure that all self-driving cars would be programmed to look for that symbol. People would then buy clothing with the symbol stitched on it or carry a flag that would display that symbol. This seems on the surface as a nifty idea, but it has numerous drawbacks and likely won’t get much traction (see my columns for discussions on that matter).

Another big topic of controversy involves what should happen when the AI driving system does detect a pedestrian. As earlier pointed out, humans will do battle over whether the car proceeds or the pedestrian proceeds. Right now, the default by most of the AI driving systems is that the pedestrian proceeds. Unfortunately, this has already led to some people deciding to leverage this default and thusly jaywalk at will (again, see my coverage).

In any case, keep in mind that the AI driving systems typically use more than just video alone to sense the driving scene. By-and-large, everyone also uses radar, and nearly all are using LIDAR too (for my analysis of Tesla’s decision to not use LIDAR, see the link here).

Various research labs and entrepreneurs have been experimenting with using specially adapted clothing and electronics to aid the radar detection of pedestrians. Pedestrians would wear the clothing, such as an adapted jacket or coat, which is lined with materials and electronics to boost the radar detection that is underway by any nearby self-driving car.

In brief, the radar sensors on the self-driving car are transmitting radar signals. Those radar signals reach the objects in the driving scene and bounce back to the radar unit. Based on the time of flight and other factors, the radar is attempting to figure out where objects are, their distance from the self-driving car, and the shape of the object (along with the movement or speed of the object).

Radar is not always going to get a good reading. There are lots of reasons that a radar signal might get stymied or distorted. In that case, if a pedestrian was wearing or holding something that could boost or aid the radar signals, it could enhance the chances of the radar detecting that the pedestrian was present.

At times, during a driving journey, the AI driving system will tend to rely more so or less so on the video camera and more so or less so on the radar. All told, the AI driving system usually has a Multi-Sensor Data Fusion (MSDF) capability that involves trying to combine and reconcile the detections coming from each of the sensors on-board the vehicle.

Another approach entails using a beaconing device that emits a signal to indicate that a pedestrian is present. The beacon could be worn by a pedestrian or possibly handheld. Clothing makers might opt to sew such a beacon into the lining of jackets or perhaps attach the beacon to a hat or cap. Self-driving cars would presumably attempt to detect the beacon signal and then accordingly take into account that a pedestrian might be nearby.

Various questions remain unresolved about the use of pedestrian beacons. For example, would the beacon merely alert that a pedestrian is within some range, or would it provide specifics such as where the person is and other parameters about the pedestrian?  Would pedestrians inadvertently become emboldened about crossing the street simply due to possessing a beacon? Would a self-driving car be expected to provide a confirmation response to the wearer of the beacon? Suppose someone placed a beacon on a fire hydrant or light pole, would this confuse the self-driving cars and undermine the value of having beacons? Etc.

Furthermore, a related issue would be whether the sensor suites of self-driving cars would already have the needed facility to detect the beacons. If not, this then raises the issue of having to include additional sensory equipment into self-driving cars, likely raising the costs associated with self-driving cars and creating other concomitant difficulties.

Conclusion

Some howl at the idea that they might need to purchase and wear some specialized clothing or have a specialized device to forewarn self-driving cars about a pedestrian being nearby.

Outrageous, these clickbait headlines are blaring.

It is a typical reaction when not looking closely at what is actually being proposed. Researchers and entrepreneurs of a responsible nature are not suggesting that you must use these products, just as the same argument can be made about wearing fluorescent jackets or having a bright orange flag. These are simply helpful means of increasing the chances of being detected.

Doomsayers are wanting to immediately decry that this suggests that self-driving cars are inherently flawed and should not be on the roadways until they can perfectly detect pedestrians at all times, in all weather conditions, in all manner of driving scenes, etc. That’s an expectation that is unrealistic and myopic.

Some are also holding out for the use of V2P (vehicle-to-pedestrian) electronic communication (this is somewhat akin to the earlier discussion about the use of electronic beacons). The notion is that a pedestrian holding their smartphone can communicate electronically with nearby self-driving cars, and so too the self-driving cars can convey messages to the pedestrian. In this manner, an approaching self-driving car might send a message to the pedestrian at the street corner and ask them to remain still and not yet enter into the street. This might help, but it certainly won’t be the panacea that some are betting on (as I’ve covered in my columns).

The reality is that pedestrians and cars are still going to be mixing together on our streets. Human-driven cars have the foibles of human drivers. Self-driving cars will be limited by the nature of the sensors being used, the AI capabilities programmed, and the driving scene conditions, along with the nature of the pedestrian in terms of where they are, what they are doing, and what they intend to do.

My prediction is that these pedestrian augmented products spurred into existence to aid in being further detected by self-driving cars will gradually and inexorably be realized as providing value.

This is still a quite nascent market.

There aren’t many self-driving cars involved in today’s tryouts. There is not much utility in selling and nor someone buying such devices until there is a sufficient volume of self-driving cars on the roadways. That being the case, we can also turn our attention to the Level 2 and Level 3 cars. They will be likely adorned with the same kinds of sensor suites as used on self-driving cars. As such, the market for these pedestrian-oriented products could arise sooner than the long-awaited emergence of true self-driving cars.

Would a doting parent consider buying a fluorescent jacket that also has a specialized radar reflective array sewn into it, doing so to aid in protecting their child from the vagaries of human drivers and the advent of self-driving cars?

One supposes if those parents are the same ones that buy smart socks for their babies (which monitor heart rate and oxygen levels, plus come with Bluetooth connectivity) and smart diapers (I won’t say what those do, you can probably guess), they are bound to want anything that can give their treasured toddler or youngster a decided advantage when crossing the street.

Mark my words, you’ll ultimately see these kinds of pedestrian wearables at your big-box retailers and even those neighborhood mom-and-pop stores.