A more realistic route to autonomous driving
economist.com
economist.com
Even outside the US, there are many areas (like UAE), which have amazing highways, sparse traffic and there isn't rail/air connectivity to some places. Self driving "shuttles" seem perfect for those applications.
It doesn't enable end-to-end autonomous taxis, so you need a sober, licensed driver in the driver's seat even if they don't need to be able to take over on a second's notice.
Handling long distance highway driving would actually be a big win. It just doesn't stir imaginations the same way that completely eliminating a driver door-to-door does--especially among people who really want to get out of car ownership and driving entirely.
I have brought this up before and I still do not understand how or why this idea is so tightly coupled to fully autonomous vehicles. I don't see how removing a human driver from a car service is going to change the dynamics enough to drive legions of people who own cars to get rid of them if they haven't done so already. The only friction a human driver adds right now is cost, but I don't see the cost savings of removing the driver being high enough to move the needle.
For any imagined scenario involving autonomous vehicles ushering in new modes of sharing, renting, etc., a useful question to ask is: "Why don't we have this today?" And, if the answer is anything other than "$10/hour for a human driver makes it too costly," then you should probably reconsider.
Do you have a source on this? It’s my understanding that most mass transit systems in the US spend a huge amount on operator compensation. BART, the one I’m most familiar with, spends $500m of its $691m operating budget on salary and benefits [0] despite the fact that BART trains already drive themselves!
[0]: https://www.bart.gov/sites/default/files/docs/FY18_Budget_Su...
Second, that $500mm covers every employee of BART. The administrators, maintenance personnel, police, and other non-operator functions do not go away. Without knowing what portion of that line item goes to operators one cannot make a judgement on how much cheaper the system would be without them.
All that said, trains are basically the most efficient mode of transport from a (number of drivers) / (number of passengers) perspective.
But speaking seriously for a moment:
- Lyft charges riders about $2/mile in the Bay Area [0] before fees.
- The IRS lets you depreciate your vehicle at a rate of $0.545/mile.
- A reasonably efficient car burns ~40 mpg on freeway, at $3.50/gallon that's ~$0.09/mile.
So, of the $2/mile the riders pay, about 1/3 of that is costs associated with purchasing and operating the vehicle (according to IRS estimates) and fuel.
If we eliminate the rest via autonomous driving, and taxis became 65% cheaper, I suspect many people would make much more use of them.
The most compelling thing I can pull from your comment is "autonomous cars will be 3x as expensive as regular cars (at least), so you won't save anything". That may be short-term true, but is almost certainly long-term false.
A 50% utilization compared with 100% utilization makes almost no difference in vehicle cost structure -- if anything, it actually increases the labor fraction of taxi cost.
If you're driving your car 50% of the time, then wear and milage is the dominant depreciation factor (rather than age). But a human sitting around doing nothing but waiting for that 50% of the time doubles the labor (time) cost of the service, suggesting an even greater savings from eliminating the human driver.
What am I missing?
The machine's time cost is time-based depreciation, which is pretty low.
I couldn't find a breakdown in the paper.
So yes, it does lessen the cost of renting cars. But it doesn't make it anywhere near too cheap to meter.
An auto-car just sitting costs almost nothing per hour (capital cost, but very little other cost), so you can have lots of them everywhere. Pickup time can be very short. I almost never used taxis in my town because of long and unreliable pickup times. Use uber all the time now. I still own a car, but if I did not have off street parking like many in my neighborhood, getting rid of it would be a no-brainer.
I'm not sure it's fair to dismiss capital cost parenthetically as part of "almost nothing". Maybe if the auto-cars were retrofitted 10-15 year old vehicles and the retrofitting cost a fraction of the vehicle's value, I'd agree. However, if they're all brand new (and even electric, where the TCO is weighted toward capital and away from operating costs), I say it's far from almost nothing.
Additionally, you go on to point out:
> if I did not have off street parking like many in my neighborhood, getting rid of it would be a no-brainer.
Therein lies a cost of "just sitting" that is routinely brought up in any car vs public transit (or other alternatives like bicycles or walkable designs) conversation.
The denser the area, the higher the cost. I suspect that if you did not have off street parking due to density, the parking portion of the "just sitting there" cost included in the pricing of the auto-car would make it no longer a no-brainer.
Where auto-cars could at least gain some space efficiency is in being able to park bumper-to-bumper and with minimum side clearance, since there would be no need for a human to enter or exit while it's in storage.
Maybe if interest rates go up capital costs will be an issue, but using a 100k car paying 4% interest is $4000 a year or about $0.50 per hour. When not moving there is basically no wear and tear or fuel costs. Almost nothing compared to paying a driver.
Driving away is, of course, not standing still. That increases the cost, but, perhaps more importantly, increased latency (or, rather, jitter).
They need not be stored in the most expensive, densest areas, but, for those who want to get rid of car ownership, that non-zero cost (especially if borne by the public with free on-street parking) isn't likely one they're willing to ignore.
> about $0.50 per hour.
A car sitting idle for 9 hours per day and averages 30mph while driving adds 1 cent per mile. A car sitting idle for 18 hours adds 5 cents per mile.
The former may be almost nothing, but the latter isn't.
> Almost nothing compared to paying a driver.
I still disagree, but that's just semantics.
If interest + depreciation is about $9K/yr, that's $25 a day, which means if you drive 8 hrs a day and have a 58% utilization rate, is $5.38/hr which is not all that far from what Uber drivers net. At 30 mph, it's 18 cents added per mile.
A self-driving car could perhaps operate more than 8 hours a day, but that doesn't mean that customers will be equally available at all hours.
I strongly suspect that when you do a proper analysis, self-driving cars are just competitive and not hugely cheaper.
That's tough to do, since we don't know either one. I suspect there's also a (perhaps incorrect) assumption that an unused auto-car will be "just sitting there" rather than moving, in which case utilization is, effectively, 100%.
That's not possible with a human driver in at least the case of a long one-way trip. However, with auto-cars, this could end up merely increasing the capital cost at the expense of unpaid/empty trip cost.
The extra cost is something I mentioned in my latency/jitter comment, wrt cheaper parking.
> $5.38/hr which is not all that far from what Uber drivers net
I didn't check your arithmetic, but, if so, it's is quite far from the fair/livable wage of $15/hr that at least I've been assuming for this conversation.
The fact that rideshare services are currently, effectively, paying much less is certainly an indication that even that may not be enough to lure people away from ownership.
It depends on what the denominator is.
If it's only the time between when a driver/auto-car "accepts" the ride and the end of the ride, there's little reason [1] to expect a drastic increase.
However, if the denominator is the total time the driver is "on duty", which is, I believe, what is generally used to calculate rideshare drivers' effective hourly compensation, then my original point stands. That is, an auto-car can be "on duty" even while just sitting in storage.
The current algorithm also doesn't tell rideshare drivers where to be while on duty, only routing them once a ride is requested. In the auto-car scenario, the computer has complete control, so a predictive algorithm could increase utilization, even if the denominator is time-in-motion.
Whether any increase would be drastic is debatable, but there's opportunity for something.
[1] Currently, the computer routing algorithm has an incentive to optimize for time at the expense of distance (since it's the driver who bears the expense of the unbilled distance, AFAIK). In the case of an auto-car, that perverse incentive would be absent, but I don't expect the difference to be huge.
If you consider the robot to "work" for more hours than a human, that's great, but due to the lower average revenue, it needs to be cheaper in order to be competitive with humans. There's no way to move your self-driving car to the opposite side of the world for the night.
"Whether any increase would be drastic is debatable, but there's opportunity for something."
If self-driving cars are cheaper, it seems like that would lead to more of them driving longer hours than humans, which would lower the utilization rather than increase it.
It's unclear to me, but this may be tautological, or at least reversible (by being cheaper to be competitive, it reduces revenue). Perhaps I'm missing your point?
> There's no way to move your self-driving car to the opposite side of the world for the night.
Ah, but that's moot. Unlike with a human driver, an auto-car doesn't need to be moved to the opposite side of the world (or an approximation, like Sacramento from SF) for the night.
> If self-driving cars are cheaper, it seems like that would lead to more of them driving longer hours than humans, which would lower the utilization rather than increase it.
You're still confusing "driving" (in motion) with "on duty".
Also, even if an auto-car is cheaper while driving, that doesn't matter if the different parties are bearing the cost of off-meter driving. In the auto-car case, it's the vendor (e.g. Uber), so there's a strong incentive to maximize utilization. In the human case, it's the driver, so the vendor has no such incentive (nor even the ability during "on duty" but not driving-for-that-vendor times).
Bicycle parking is also a problem in many areas that have a high density of bicycle users. See Amsterdam train stations or apartment buildings where you have to store your bike in your 600sqft apartment.
Walking is really the best although these new electric skateboards seems to be almost totally portable and storable.
That is significantly underestimating the operating costs of a vehicle, especially a vehicle that is used heavily. There is a lot of periodic maintenance required, and the frequency increases as usage gets more stressful. Then there is non-periodic maintenance, be it from accidents, vandalism, or just flawed parts.
It's true that most of these costs come in even if you own, but putting 200k miles on a car over eight years costs less (both per-mile and per-unit-time) than putting 200k on in two years.
The per mile part sounds like a personal speculation. (Per unit of time is obvious and irrelevant here)
Quite the contrary, it's most efficient to put the load in a short span to minimize time-limited consumables like belts and seals, fixed costs like insurance and regulatory charges per vehicle, interest, depreciation of things like paint and esthetic appeal etc. The battery has an age induced wear also, the tech is moving fast and more efficient vehicles are always appearing and so on.
Modern vehicles are incredibly more reliable than they used to be, but this seems to be overstating the case.
At $70/hr., I expect people will still opt to own their own vehicle.
The cost is upfront. People sit down and budget $x for transportation and then they know that amount of money will get them anywhere they want to go. People hate having to decide in the moment whether they want to spend the money or not. I notice a similar phenomena when it comes to mobile phone service. People would rather pay extra for a plan that provides more than they could ever need than to have a lesser plan and pay overages when needed because the overages require thinking about it in the moment, rather than pre-planning how much they want to budget. $70/hr. isn't even going to have anyone thinking twice about changing their behaviour completely.
Maybe robs-taxi services will pioneer a pay ahead of time service that captures the necessary mindshare that will make ownership obsolete, but that won't have anything to do with the technology. There is nothing stopping a taxi service from doing that today with human drivers. It's just a tough business model to work with when providing services, so it is uncommon.
I make this point every time I can: kinetic energy is the killer and it goes up with the square of velocity. Half-speed is quarter-deadly.
A golf cart-like machine, limited to a maximum speed not much faster than walking, light-weight, and festooned inside and out with airbags (and other anti-collision/-injury stuff), could be built today, I think it would work just fine, and I bet you could sell a lot of them. ("Baby Boomers" he says under his breath...)
Trying to go straight to robot cars is hubris, it's biting off more than you can chew and then trying to talk past the food in your mouth.
Also, can we please call them "auto-autos"?
https://trl.co.uk/news/prev/35671
The last mile if far more interesting at the moment, with shuttles, Ofo bikes, e-bikes and scooters and "baby boomers" on their mobility devices. What it really needs though is more infrastructure to support it, ideally dedicated lanes.
The next step would be 1h-out-of-town commuters + close-suburbia to specific dropoff points in downtown or wherever jobs tend to be. Again, stay on big roads that have good signalling, and keep the bright colors so other drivers will know to stay out of your way in the carpool lane.
It's basically what the big tech companies in the Valley are doing with their commuting busses, but smaller and for everyone. The min-busses with flexible routes are a common mode of transportation in certain Eastern European countries, see https://en.wikipedia.org/wiki/Marshrutka
They are. But that makes me wonder why we don't already see a lot of this type of thing in the US. My instinct would be the cost of a minimum wage-ish driver isn't the difference between this mode of transportation working and it not. Of course, it does exist in a sense with Lyft Line and Uber Pool although those don't get a lot of attention.
https://en.wikipedia.org/wiki/Demand_responsive_transport#Un...
Lower costs there can perhaps lead to a virtuous cycle of increased service and use.
The one here is open to anyone and even sometimes runs a loop for a local taproom.
I guess you could sample 10 or 100 of them and see how many impose restrictions on rider eligibility.
> The one here is open to anyone and even sometimes runs a loop for a local taproom.
If it is running a loop it isn't DRT, it's a jitney.
Much easier (though still very hard).
I know it is a bit of a fallacy to say "we can fly to the moon, so we should be able to have a self-driving car", but it's also not clear that the tools we have (ML, simulators, planning algorithms, etc) are not up to the tasks that have Drive.ai have engineered around.
It's a reasonable strategy to elide things you can and focus on the "core" hard task, but if solving these other difficult problems is also meaningful product-wise (and it clearly is meaningful), there's no reason you can't have people working on both. And given Waymo seems to be gearing up for a large launch, I have a hard time believing that these additional tasks are as hard as Drive.ai want to make out.
Wrong, the car did spot her, but the emergency braking systems were turned off.
The problematics of autonomous trains are vastly simpler than autonomous cars.
Trains move in restricted paths (tracks), trains should stop only at specific places, trains meet other trains in such a few possible places, and so on.
You can try to simplify the problem by limiting your routes to major highways and making the vehicle ultra-visible so other drivers give you a wide berth.
But at some point there will be construction, or a car wreck, or debris in the roadway. Eventually you'll need to exit the highway and drive on city streets.
Even if you manage to reduce the hard stuff to 1 mile out of 200, you still need to deal with it.
If you can't handle everything, you're not safe to be on the road.
I think the minimum required to deal with it is to have the car pull over when it identifies a situation it can't handle. Even an onramp to exit car that pulls over if it sees construction will be extremely valuable to anyone routinely doing highway driving.
But I wouldn't be too quick to interpret a single sentence from a piece of short-form popular journalism as a precise specification of what their cars are and are not designed to handle.
However, it is exactly as it should be. I applaud them for daring to be dumb and focusing on the easy parts of the problem first. I think this will catch on and get better slowly.
The manufacturing learning curve has been the path to lower costs through mass production and tunnels will be no exception.
“Without tunnels we’ll be in traffic hell forever” -Elon Musk
You'd only want to bring a surface car for trips with rural or suburban legs on BOTH ends. Which seems relatively rare.
[1] https://www.boringcompany.com/
[2] dynamic routing means no waiting around to change trains
I've seen that Dr Who episode.
There are plenty of air traffic regulations to deal with
Gasoline consumption is not going up the way it used to, and there apparently is a whole lot more oil than a lot of people thought-- I'm guessing due to a combination of countries not providing correct numbers (it's a national security issue after all), new oilfield discoveries and new extraction techniques. Add in more efficient vehicles and alternative energy sources, and you quickly realize the worst that's going to happen is some gradually increased prices for products & services most closely tied to the current cost of fuel.