At a grand spectacle in San Francisco last night, the GM-backed Cruise that aims to bring forth autonomous vehicles with true self-driving capabilities had an unveiling of their newest creation (said to have been devised via engineers from GM, Cruise, and Honda).
It has been christened as the Origin.
Looking akin to a minivan, the Origin is a boxy vehicle touted as being an EV that is destined for ride-sharing purposes, and was feted as a solution that is self-driven, all-electric, and shared.
The evening event was primarily devoted to the hardware side of things, eschewing any in-depth indication about advances on the AI and software side (actually, there weren’t many specific details on the hardware either, though an actual vehicle was presented and made available to touch and see).
Alas, there were no details about the battery range, nor what kinds of speeds are intended, nor maneuverability characteristics, and most importantly for those focused on self-driving tech there wasn’t a detailed indication about the sensors, sensor types and what their capabilities consist of (more on this in a moment herein).
You might rate this as an unveiling with perhaps one eye closed.
That being said, please don’t misunderstand that somehow showcasing the nature of an intended self-driving car vehicle and not covering all the rest is necessarily an undue affair.
We are going to need a physical car or some kind of vehicle for self-driving transport, so the hardware in terms of the overall shape and nature of the automobile is indeed important (for example, see my posting about MAVS, Modular Autonomous Vehicle Systems).
It is helpful to the industry and the advancement of self-driving cars to have the automakers strut their stuff in terms of what they envision true self-driving vehicles are going to be like (for my prior coverage about Cruise, see the link here).
Some immediately expressed that the Origin seemed to resemble the VW Cedric concept self-driving van that was unveiled in 2017 at the Geneva Motor Show.
Well, come to think of it, there is such a resemblance, along with similarities to many other concept self-driving vehicle box-designs that have been floated over the years.
The point is that we are gradually witnessing a coalescing toward one kind of specialty of a self-driving car, essentially a pod-on-wheels (which is different from a box-on-wheels, a type of self-driving car used primarily for delivery of goods).
Does this mean that all self-driving passenger cars are destined to be boxy pods?
Let me say that again, no.
It just means that of the various styles and designs for self-driving cars, one that is going to gain traction is the pod.
Indeed, for the cynics that were quick to challenge the pod-like design of the Origin, you’ve got to keep in mind that customarily a vehicle is purpose-devised (structured and shaped for a specific intended purpose or use case).
If you want a vehicle to serve a particular market or purpose, it makes sense to design the vehicle accordingly.
Sports cars look like sports cars, while minivans look like minivans, and you can’t carp about the matter.
Of course, you could try to criticize if someone was aiming at sports cars though the market needed minivans, and likewise, you could complain if an automaker pursued minivans when instead the market really needed sports cars.
In the case of the Origin, since the stated purpose is for ride-sharing, the use of a pod-like design is in keeping with that purpose.
This does not preclude other kinds of self-driving vehicles to also emerge, and nor does it suggest that we will only have self-driving pods on our roadways.
One slightly over-the-top notion though is the suggestion by some that this is a reinventing of what we think a car is.
Sorry, a bit of an overreach.
I’d argue that we are still going to have “cars” in terms of genuinely looking cars that are self-driving. In addition, we’ll have pods or minivans that are self-driving. We’ll have buses that are self-driving. And, I’ll blow your mind perhaps by claiming that we’ll have sports cars that are self-driving (see my discussion at the link here).
The Origin is essentially a shuttle or pod or minivan, whichever vernacular you prefer, and represents one segment of the burgeoning self-driving “car” marketplace.
Now that I’ve cleared up that aspect, let’s get to the brass tacks.
Based on the rather scant reveal, and subject to further elocution once more details are inexorably shared, here are three key ways that the Origin might find itself facing some heartburn.
This is not solely aimed at the Origin, and other pod designs are likely to have similar kinds of both strengths and weaknesses that need to be contended with.
Interior Space Design
Imagine that you could chuck out the window of a conventional car the steering wheel and pedals.
Man, you’d have a great opportunity to rethink the design of the interior space.
There’s no need to have a seat devoted to a human driver.
You can put the seats wherever you want.
One of the most popular concept designs consists of swivel seats, allowing passengers to swivel and face each other, or face toward the front, or face toward the back, or face toward the sides and look out the windows in whatever direction they prefer.
Furthermore, the swivel seats allow for a potential working arrangement, whereby you might place a small table at the center of the vehicle. Passengers could face each other, carry on a dialogue, and use the table for getting their joint efforts undertaken.
That’s not the Origin.
Another concept design consists of having seats that are luxurious and recline.
In fact, the reclining might be so extensive that you can pretty much lay down and sleep while inside the self-driving vehicle. For those that take long commutes to work, it could be a godsend that you’d now be able to catch some snoozing, so you’ll be fresh and ready to get to work once you arrive at the office.
That’s not the Origin.
For the Origin, the designers apparently decided that this pod would be for people on-the-go that are going to leap into the minivan, be whisked to their destination, and then jump out once they get there.
No sleeping (presumably), no working (at least not in the swivel seat manner), all of which makes sense to not do if you are on a quick-ride shuttle.
Here’s how the interior is shaped.
There are two benches, essentially, each bench facing the other, akin to if you sat in a shuttle.
One bench faces toward the back, the other faces toward the front.
Each bench consists of three seating positions, though it is really more like two people and maybe you could jam in-between them a child or someone rather slim in stature.
The seating capacity is comfortably for four people, and the stated capacity is six people, though as I say you’d be somewhat hard-pressed with six strangers sitting in there (they’d get to know each, a bit more so than they might prefer, if you know what I mean).
The seats appear to be unadorned, meaning that it has that straight-backed feel as per a shuttle, unlike bucket seats that you’d find inside an everyday car.
Generally, you would not likely wish to sit in such seats for any lengthy trip, and thus once again reinforces the design notion that this is a shuttle or pod for short hops along with a quick entry and exit.
Under each seat is a small space for passengers to place some belongings, similar somewhat to how you might put items under your seat while in an airplane, though in this pod you have ready access to the area directly underneath your own seat.
Here are a few heartburn aspects of the interior space design that I predict will arise once the Origin is used in real-world settings:
· Placing items under the seat
Riders are going to forget that they placed something under their seat as they hastily exit from the Origin.
Meanwhile, presumably, the Origin scoots along to whatever its next destination is.
The passenger that left their item might later realize they mistakenly did so and will try to access the ride-sharing network to see if they can get that Origin to come back to them.
That’s a logistic nightmare.
Meanwhile, perhaps someone else has already taken the item, figuring that it’s a dog-eat-dog world of winners are keepers and losers are weepers.
You might normally have had a chance that a human driver would realize a passenger left an item, and perhaps keep it secure or report it to HQ but recall that there’s no human driver involved.
If there are cameras pointed inward in the Origin, which there might well be, and its something that many self-driving vehicles are going to include, the ride-sharing system could potentially take a peek to see if the item is still there, though this raises other privacy issues and isn’t a slam dunk solution.
They might put sensors on the floor under the seats, trying to detect the presence of objects, and then alert a passenger as they get up to leave, though this too has a slew of complications.
Anyway, as you can see, a driverless shuttle with the nicety of under-the-seat storage has its trade-offs.
· Storage at the front and rear
Perhaps even more vexing is the storage compartments that are apparently at the front and rear of the Origin.
It’s a good idea to allow people to place their bulky items in such storage compartments, aiming to prevent clutter inside where the passengers are sitting, but unfortunately this is a whole can of worms.
Assume that you are in a rush and want to use an Origin to get a few blocks down the street in your downtown area.
An Origin comes up to the curb, you get in and buckle-up.
The Origin goes about a half-block and picks up the next passenger.
This person has two bags of groceries and wants to place them into the front compartment. They fumble with the compartment and feebly place their bags into it. Oops, one bag tears and the contents spill. The person tries to pick them up and shove them into the other bag.
All of this is taking time.
You are seated in the Origin, wanting to get a few blocks down the street, and your ride is suddenly going to take ten times longer than you estimated.
If there was a human driver, perhaps the driver would help the person load the grocery bags into the trunk and having done so zillions of times the driver is very adept and quick at doing so.
Again, there’s no human driver and so the passengers are on their own.
Any self-driving shuttle or pod that allows for compartment space is going to ultimately have to deal with these kinds of issues.
Do you restrict that no one can use the compartments?
The Same Experience Every Time Is Not Realistic
It was indicated at the unveiling that the Origin will not be a car that you can buy.
Instead, it will be presumably owned by the automaker and deployed as a fleet of self-driving cars, exclusively used for ride-sharing purposes.
That’s fine if that’s the market position they wish to take.
As I’ve stated repeatedly, I believe that there will nonetheless also be a market for private ownership of self-driving cars, though I’m a bit of a contrarian since many pundits claim that self-driving cars will only be owned by large firms and used solely as ride-sharing fleets (for my explanation about why private ownership is also viable, see the link here).
In any case, the beef that I have about the matter is the claim that supposedly every time you ride in an Origin the experience will be the same every time.
This was explained by the example of getting into an Uber or Lyft today and being knocked over by the smell of Doritos and Mountain Dew, or the biting odor of over-scented Pine-Sol, all of which presumably occur because of a human driver that’s either obnoxious or oblivious to the stench inside their ride-sharing car.
In theory, take out the human driver, and you won’t have any Doritos and no Mountain Dew.
But, wait, think about that.
You have passengers.
Joe gets into the Origin and has a liverwurst sandwich.
Jane also gets into the Origin and has a raw onion that she loves to eat just like one would eat an apple.
Sam also boards the Origin and he’s got a large bag of Doritos Nacho Cheese tortilla chips and his trusty Big Gulp of Mountain Dew.
They have quite a fun time during their ride, and part of Joe’s liverwurst sandwich ends-up on the floor, while Sam accidentally spills his Mountain Dew onto the seat, and Jane leaves onion peels all over the place.
Upon reaching their destination, out they go.
You were waiting for the Origin and happily step into it.
Yikes, you can barely breathe!
As I’ve repeatedly stated, a looming problem for the self-driving car ride-sharing era will be the stench fest that is going to arise.
This could be ten times worse than when you have a human driver since the human driver is likely to notice odors or at least do something about odors once they realize that customers are upset.
There’s no human driver to clean things up or try to deal with foul odors.
Passengers are free to do whatever they like.
Sure, you might say that the fleet owner needs to ensure that the inside is clean and odor-free.
But, how do you do that for each and every ride?
You can’t have your self-driving vehicle continually coming to the home base to get cleaned upon each ride given, it just doesn’t make any viable sense.
So, for clarity, a self-driving car that’s being used for ride-sharing is not a guarantee of the same experience every time, other than for the driving aspects and the nature of the vehicle (assuming that a fleet is only using one and only one type of vehicle), but not for the nature of the interior and aspects such as odors, cleanliness, etc. (for the potential of e-nose sensors, see my posting here).
A new sensor was showcased at the unveiling of the Origin.
No specifications were provided, but it appeared to be a sensor that is intended to sit at the exterior corners of the vehicle and pivots back-and-forth on its mounting as though scanning what is in front of the sensor.
This was likened to an owl that moves its head.
Or, akin to how humans move their heads, scanning back-and-forth to survey the driving scene and gauge what’s up ahead and to their sides.
The analogy to animals and humans seems compelling.
Sure, if via Darwin’s theory we and animals evolved to move our heads back-and-forth, doing so to try and observe both friend and foe, it must make sense to do the same thing in a car.
Anytime that you have a sensory device that relies on physical movement, you are asking for trouble.
The device is likely to wear out or breakdown as a result of the ongoing and presumably continuous movement.
If you rightfully assume that ride-sharing self-driving cars are going to be active 24×7, which makes sense to get your monies worth, other than brief windows of time for maintenance and upkeep, this means those pivoting sensors are going to be on-the-go a lot.
Let’s hope the failure rate is really low.
That’s partially why too there has been a shift in LIDAR field, another type of sensor often used on self-driving cars, involving reducing the number and nature of the moving parts.
Setting aside that concern, another involves whether the pivoting scanner will be looking in the right direction at the right moment in time.
We’ll go back to the human analogy.
You look to your left, the coast is clear, you look to your right, and start to make your left turn. Oops, in the moment that you checked the right side, a bicyclist suddenly appeared to your left. Bam, you ram into the bicyclist.
The question is how fast the sensors pivot and whether the pivoting potentially opens any untimely gaps of not spotting what’s going on.
There’s also the whole aspect of sensor fusion too.
How is the pivoting sensor stitching together the data collected at each stance of the pivot?
Does the pivoting sensor on the left side of the vehicle get mated with the pivoting sensor on the right side of the vehicle, crafting a cohesive whole?
I’m certainly in favor of new kinds of sensors for self-driving cars and have argued vehemently that the more sensors the better, which is contrary to the views of some pundits, though I’ve also said that not all sensors are ready for prime time and it’s important to make sure that the sensors perform a needed function, each providing added value, and collectively working on a unified basis.
Time will tell.
Kudos to GM/Cruise for their willingness to showcase their latest approach to a self-driving car.
Some self-driving tech is being added to conventional cars, such as Waymo’s efforts with the Chrysler Pacifica, while in other cases an entirely new vehicle is being made to support self-driving.
There are trade-offs whichever way you go (see my link here).
At least we know this, the more that the automakers bring forth their efforts, the sooner that the nature of self-driving cars will evolve and mature.
You might say it’s the origin of a new species.