You are here

Sleeper cars

Volvo concept in bed mode

Yesterday, Volvo got some good buzz for a concept car which included a bed for sleeping and asking the question "why fly when you can be driven?". I've written about sleeper cars before, as well as the full robo-RV, but let's put all the issues together.

Obviously, if you can sleep while you travel, it's magic, because you can make the travel take no time at all. Those who sleep easily on planes (not me) actually view red-eyes as their preferred flights. There is the potential for the sleeper car to be the cheapest and most efficient vehicle on the highway, and create a whole new style of travel.

  • Compared to a plane (and even a train) space can be at less of a premium.
  • If you have a car which is only a sleeping car, you can actually make it super low to the ground, which can give it much less drag, for more efficiency. In fact, the solo sleeper car should be the most efficient passenger vehicle on the highway.
  • If the passenger wants a full night's sleep, the car can prefer to go at a slower speed (as long as that doesn't impede traffic) for even more efficiency.
  • Like all robocars, with no need to have the driver "feel the road" one can use a dynamic, super-soft suspension to make the ride more comfortable, though perhaps not as smooth as a plane with no turbulence or a high quality rail line.
  • The problem of seatbelts and airbags for a sleeping passenger may be more challenging than for a sitting one. While accidents are more rare, they will happen.
  • For overnight trips, a place to change clothes and do morning ablutions is also needed.

There are a few situations where people might like to sleep:

  1. Overnight trip to city 250 to 500 miles distant
  2. Family trip to vacation home. Load family Friday night, arrive Saturday morning. Load family again Sunday night -- more total wake-time at destination than day-driving.
  3. Commuter with long commute sleeps less at home, completes sleep on the way in. Possibly also naps on return.
  4. "Nap" car -- for those who get drowsy on any trip.

The latter two choices are easier in a car that converts -- ie. a seat reclines to a bed. They can also be done in a dedicated sleeper car if, once you waken, the car pulls over to meet up with a normal robotaxi with upright sitting. A family sleeper needs a toilet, which makes it a fair bit bigger. Large family and group sleepers could have two levels of bunks.

The morning commuter car, good for 1-2 hours of sleep, is the easiest. It doesn't need a toilet or place to change. By making 2 hour commutes tolerable, it would increase sprawl and exurb development, and waste energy. Group commuter cars, which bunk 4-8 or even more people and use single person pods to bring solo passengers to and from them, could provide an efficient option. Pity the person who decides after boarding that they can't sleep.

Sleeping on the morning commute is an obvious thing. It's a little more difficult on the return for most. Employers don't want you in "needing to sleep" mode in the last few hours of work. However, the natural human sleep cycle actually is an early rise and a post-lunch siesta nap. It might be possible to adjust this to the commute regimen.

Of course, it's always possible to make vehicles like the Volvo, which convert and have both nice seating and a bed, and enough space and height for a toilet or working -- like fancy first class seats on overseas planes. But this comes at a serious cost, both in the price of the vehicle and the energy use.

Why the focus on efficiency?

In spite of what Volvo says, trips of 400 miles in group vehicles, like planes or trains (or buses as I'll describe) are more efficient than a solo trip in a larger and heavier vehicle like their concept. However, it's right on the line, and so if the vehicle can be low-slung or slow, it's going to beat the plane. And it's always better to use less energy.

This is even easier if the vehicle is electric. While having a 500 mile battery (which weighs a lot) may not be the right choice, if the drive takes less than 8 hours, a sleeper car can pull off (using very gentle and planned stops) at charging stations designed for a recharge during the night. An electric sleeper will beat the airplane for energy use.

The low-slung solo/couple sleeper might be the cheapest highway-capable car there is. In the electric design, it would feature just a base, 4 wheels, suspension, a battery pack with mattress on top, roll-bars and crumple structures, airbags and restraints and a simple fiberglass dome roof that hinges on one side to lift up. Perhaps a small window for comfort but possibly not even that. No big screen, big speakers, buttons or knobs or adjustments. Charging for your devices, small speakers and a small screen and an emergency exit and that's about it. Even the ICE version is simple with a small engine (no need for fast acceleration or power unless it's doing the mountain routes.) With no chair and no need for things to occupy you, it's the cheapest vehicle on the road. And if it's just 3' wide (solo) or 5' wide (couple) and slung low, it's also the lowest drag vehicle on the road. Adding an aisle and some height to bet to the toilet makes it bigger and more expensive. Making it an RV probably makes it 4x to 5x more expensive. Which means the solo ride costs that much more too.

Follow-up on efficiency of solo transport

I write a follow up article on the surprising efficiency of solo transport when it comes to sleepers. It provides more comment on this and other issues.

The toilet

For overnight trips, the need for a toilet provides all sorts of challenges. That's not practical in a low-slung bed-only car. While some (men in particular) can use a bottle lying in bed, it's nobody's preference. Those who always go the whole night without need to pee could handle a car with no toilet (or a bottle for rare urgent needs.)

Around 55% of people don't routinely wake up during the night to pee, but the rest do it 1 or more times per night.

Of course, our highways today are full of rest stops because regular road-trippers regularly need the toilet, and so your sleeper car can pull over, though again, having to go outside for the joys of a public toilet is rarely a first choice. And nobody would like waking up with the need to go, and then learning it's 10 minutes to the rest stop. You can't go back to sleep with that pressure so your sleep is more disrupted. Those with a predictable pattern can program the car to stop and wake them when it's almost time, which can solve this problem.

Family and group sleeper cars also will run into trouble, because people have needs at different times, and parents don't want kids going out on their own to public toilets in the middle of the night.

An on-board toilet and small sink can of course be done, and the car can travel after the trip to a depot for draining and cleaning. The main issue is it costs a lot of space and height -- you have to be able to get out of bed, get to it and then sit or kneel as need be. Suddenly the vehicle needs an aisle, and more height -- it's a large cost to get this, but many people will insist on it.

For those willing to pay -- which might be many because I think having an on-board toilet might double the cost of the ride -- the roadside rest-stops could be high-end, where the electric car actually drives into the bathroom and a garage door closes behind you to allow you a private visit.

A toilet bot?

One could also imagine a vehicle whose only function is to be a roving toilet. They would not go fast or far, but when you indicated the need for one, it should probably be able to meet you along the shoulder (in a safe section, possibly added for this.) Your car and it would dock touching one another, so you just step out of bed right into it. Every so often, they would head off for draining and cleaning and return to the popular late night routes. If sleeper cars were standardized enough, they could even "dock" in an enclosed way, so you don't have to get exposed to the cold too much. These would even be popular with daytime drivers and human drivers. They would be only for urine. For #2 use a regular rest stop.

Slow cars

For a family car to a vacation home 4 hours away, you want the car to go as slow as possible. In fact, you would be so slow that this would annoy other people who are in a hurry. Of course, with a vacation home, the early arriving car can park and people can wake up whenever they like, even after a full night of sleep.

Slow cars on the road would be polite -- always pulling over if somebody comes up behind, and possibly sticking to minor roads, though they will prefer not to have lots of stops which disturb their passengers.

One radical idea is that, late at night, cars could pull over to the left, which is to say into the oncoming traffic, to allow a fast car behind them to pass. With a map of the geometry of the road, and the ability to detect oncoming headlights at a long distance, slow cars could know when it was safe to do so. They would even slow a bit while doing this. The fast car would pass quickly. The sleeper car could even flash a signal to the fast car that it is going to do this, allowing the fast car to zoom through as if the slow car wasn't even there. At night, traffic is light enough that this should work. On the highway, sleeper cars would of course keep to the slow lane.

Group sleeper vehicles

One very efficient vehicle would be the group sleeper -- effectively a bus full of bunks. Everybody wants to go overnight, and adding a shared bathroom, and even an aisle is not as much of a burden. (It should be noted, though, that otherwise a vehicle can have no aisle because people can have doors to the outside for their bunks.)

In fact, this already exists -- the group sleeper is efficient enough that the cost of the driver is not as big an issue. A service called Cabin drives between SF and LA with bunks and even a small kitchen. When it gets to its destination, it sits in the parking lot for 2 hours for those who wish to complete 8 hours of sleep.

This practice is also common in some places on trains, and on the Baltic in ships. It's particularly popular to travel from Helsinki and Tallinn to Stockholm on these ships. The price is attractive if it saves you a night in a hotel room, and they also offer meals, gambling, duty free and some sightseeing for the tourists.

At the same time, it is important to understand the vast difference between a solo sleeper and a group one in convenience. A solo sleeper can take you from your driveway to your staging destination on your sleep schedule. When you get tired you would do everything you normally do to get ready for bed, put on your PJs and instead of going to the bed, walk to the driveway and climb into bed there. If you are one of the lucky people who falls asleep quickly, next thing you know your vehicle is parked inside your staging destination (a place where you can shower and change.)

With a group sleeper, you must fit your night to their schedule. You must travel (presumably not in a bed) to the departure point, and get there with enough margin to be sure you make it. All of this is probably done awake. You could easily see the group sleeper involving 10 to 100 times as much awake travel as the private one, and only awake travel time counts for this type of customer.

Compared to the airline

It's early for Cabin, but I don't think it's a big threat to the airlines. Volvo points out the nightmare that short-haul flights have become, with 3 hours of travel with only 1 hour of flying. The "teleport" aspect of sleeping can be attractive then, except for two things.

  1. While you might want to be there in the morning, the teleport illusion only works if you also want to stay until late in the night as well.
  2. The airline nightmare is fixable, if we have the will -- we have done this to ourselves. If the airlines started facing real competition from overnight sleepers, they will adapt, because they have to.

Vehicle design

As noted, the single person sleeper, just 3' wide and 8' long and 4' high, could be the cheapest and most efficient vehicle on the highway. Here are some possible layouts:

  • Couple's car, 5' wide bed
  • Family car with 5' wide bed, and transverse berths (only 5' long) for children under 4'10" in height
  • Family or group car with 2 levels of adult berths for 4, 8 or more passengers.
  • Wider solo car with diagonal bed for more safety and two triangular cargo spaces.

However, in need of more research are the best designs for safety. While the best aerodynamics are higher at the front and coming to a point at the back (teardrop shape) it's probably not good to be riding headfirst, at least in forward impacts which are the worst of all.

It's also hard to figure how to restrain people lying feet forward. No matter what direction, one idea might be a webbing which can descend from the ceiling after bedding down, and then tighten hard before an accident.

Transverse bunks are probably much better for restraints in forward and rear accidents. Side impacts are not very likely on the highway. However, for adults, that requires a car almost 7' wide which is bulky on the road and low efficiency. It does make having toilet space easier. Cars could actually come in different sizes and you would get one with bunks big enough for you (or larger.) Robocars can handle being 7' wide as they will remain reliably centered in their lanes -- it's just a question of drag. Airbags also will work better with a passenger perpendicular to the direction of travel.

To get more advanced, a bed (particularly a transverse one) could be on actuators which could quickly tilt the mattress almost vertically. You would then deal with the impact by being slammed hard into the mattress, spreading the shock over your whole body and deep into the mattress. The mattress might be extra thick. Combined with crumple zones the severity of impact could be much less.

Question for readers

An unscientific survey: If taking an overnight sleeper ride, would you insist on a comfortable toilet/urinal, if that ride cost twice as much as a car with only a bunk and an emergency bottle? Indicate if you are male or female.

Answer this question via this poll page. Not in the comments unless you wish to add more thoughts.


Mostly I'd be happy if a car had a comfortable seat that fully reclined and could pull over at a rest stop within 20 minutes if I asked it to. A bathroom is unnecessary. I think a dedicated sleeper car arrangement probably won't work, as it means the car doesn't have a function during much of the day and can only function doing long hauls at night.

Having no function during the day is no issue. This car is going to do 400 miles of driving or more every day. That's a lot more than typical cars do today, even urban taxis which work 24 hours rarely get that much. (Average New York Taxi about 180 miles/day.) In addition, it may get to add one sleeper morning (and evening) commute depending on schedules.

As a person who usually has to go during the night, I would not find a 20 minute wait to pee acceptable. Once the need to pee wakes you up, you want to go soon, and you can't go back to sleep for 20 minutes. So it means you get "fully" awake. Normally, in your house, you just peek quickly and never fully awaken. It's a big difference. Even going outside, particularly if it's cold, can be an issue.

I definitely require some provision for peeing, but it need not be elegant.

If a public toilet ever copulated with my car, I'd have to throw out my car and get a new one.
Otherwise this sounds great.
We can take care of business in Starbucks, just like we do now.

I own and drive a luxurious but inconspicuous high roof van. On the inside it's a tiny house, on the outside a work van. When I travel I drive 2-3 hours, sleep just about anywhere for 2-3 hours, and then repeat as necessary. Driving on interstates in the middle of the night there is little traffic, and I have a device to watch and interact with as desired while driving. The bathroom is simply cracking the sliding door on the side and peeing out that crack. Trips of up to 500 miles a night are doable out west where the speed limits are 80 and up. And when I get there, i've brought my hotel with me, can park almost anywhere, and perhaps just need an extra nap on the day of arrival. I shower and change at public rec centers, usually about $5. I look forward to when this is self driving, but it won't be that big a game changer compared to what I do now.

Hey Brad, thanks for the mention (this is Tom from Cabin). We think the Group sleeper (moving hotel) is the best way to deliver this experience in a pre-AV world (similar to how DVDs-in-the-mail were the best method for Netflix to deliver content pre-streaming from 1997-2007), but I agree with you that solo sleeper pods are the best and most cost effective solution in the long-term. It will require the build-out of key infrastructure along the route (e.g. high-end rest-stops every 5-10 miles that are safe) and at the end-points for the morning wake-up experience ( utilizing under capacity hotels is a good interim solution).

One thing to note about soft suspension systems is that a lot of the worst bumps on the highway (that wake people up) are at low frequency (1.5-2.2Hz). We've transported a little over 10,000 people at this point and roughly 15-20% have a tough time sleeping (and you have to think that we have a self-selecting group of customers that believe they could sleep well on a moving vehicle). The human body experiences whole body vibrations differently depending on the axis of vibration and the same exposure to vibrations you can withstand for 2.5 hours seated, you can only tolerate for about 5 minutes laying down (, page 9).

We built an active secondary suspension system to address this ( and think that most sleeper cars will need significantly upgraded suspension systems to allow the general population to sleep well. One cool thing is that you can use the active suspension system to rock people to sleep ( or back to sleep if they wake up and go to the bathroom.

I have been curious about what should be possible in active suspension for this and other robocar applications for over a decade and am very interested to see your work on it. (Sadly, Cabin would not be convenient for my trips to LA, since I would have to backtrack north 50 minutes to get on your bus in SF.) This is one of the challenges of a bus approach.

I believe it might be possible to use highly reactive suspensions such as voice coil or certain others to keep the sleeper close to motionless, even with fairly sharp shocks.

I have noticed Cabin has only 2' high berths, which I think would be a bit claustrophobic, so I wonder how people react. I also wonder how you "seat" belt your passengers. You place them on the upper deck -- do you really need all that lounge space below, and doesn't that subject them to more vibration up top?

I agree on the limitations of the high capacity group sleeper vehicle requiring centralization and redistribution at the two endpoints (sometimes one if the endpoint is a destination like Disneyland in Anaheim or Northstar in Lake Tahoe). Long term, just as Netflix switched from DVDs to streaming when the technology/economics made sense, Cabin will likely also use smaller group vehicles (6-7 passengers in a <22' vehicle) when the tech is there/commercially ready. The economics of going small don't make sense right now though because the fully-loaded driver cost is around $380 (hourly wage, benefits, workers comp, etc, etc).

In the meantime, the high capacity group sleeper solution still makes sense on a lot of routes in the US and especially for short duration trips (1-3 days) where it's important to maximize time at your destination relative to utility time spent in transit. As an example, very few people that don't own/lease a private jet or can afford a surf-air subscription could justify the time/cost/hassle of going to/from Santa Barbara or Disneyland for 12-16 hours.

One major advantage of large format vehicles over air though is that they do not require a runway and therefore can have more discreet pick-up/drop-off locations and direct routes (as you mention in your next article on sleepers). The unit size (24 passengers) is also much smaller than a 737-800 (162-189 passengers) and for a route like SF/LA where there is a flight leaving every 6-7 minutes, you can afford to have a lot of discreet direct routes (e.g. SoMA to Santa Monica, and Mountain View to Downtown LA).

We've learned a lot about the design of group sleeper vehicles from our first three prototypes and as you point out, we've learned that we do not need the lounge space below (only really used during boarding/arrival). Our next gen vehicles will have larger berths (more headroom, width, and length) as well as the active suspension tech we developed to reduce unwanted vibration and input vibration that helps people fall/stay asleep. Sleeper Berths ( have been in the US Federal vehicle code since 1953 and the safety has been well studied (the most common use of sleeper berths right now is semi truck tandem teams).

The single largest experiential benefit of the group sleeper is the economics of having a nice on-board bathroom (square footage it takes up as well as cleaning/servicing is amortized over more passengers). It's a lot harder to fit 6-7 berths into a 22' vehicle with a bathroom when you still need to have a driver seat. Once you go beyond 22' or more than 7 passengers, the design of the vehicle must accommodate 1-2 ADA tie-down(s) per vehicle for a fixed route which greatly increases its Capex/operational complexity (Chariot was fined for this-- Having a demand-responsive service, Uber solves for this with UberXL (allows for 6 passengers) and UberWAV.

Accessibility is something that is now and will always be REALLY important to us (Laws are made for reason!) and we made sure that each Cabin vehicle has 2 ADA tie-downs and that the sleeping berth on the lower floor is accessible if the passenger wants to sleep in a lay-flat instead of being in a wheeled mobility device during the trip. We tried to integrate an ADA accessible bathroom into the design of the group sleeper vehicle, but it wouldn't fit in a double decker format (so we instruct our driver to make stops if a passenger with a wheeled mobility device needs to use the restroom).

For active suspension, we studied using voice coil motors (first instinct because of speed of actuation), but their lack of mass production makes them a speciality component (high capex) relative to off-the-shelf rotational motors which can be just as responsive when used in direct drive configuration without gearing.

Would be happy to cover your Uber to SF for you to try Cabin for a weekend trip to Santa Monica. I think trying out the service could be insightful in understanding how sleeper cars will play into solutions for the future mobility stack (lots of our team's insights have come from taking the service hundred of times).


Sadly, it's not the cost of an Uber up to the city that stands in the way of Cabin being convenient for me (at least as a customer rather than a reviewer.) When you add an hour of travel in the wrong direction and the need to get there well before boarding to get a good berth, and the fact that I am a light sleeper, and that the airlines upgrade me to 1st class etc. Cabin can't compete with the plane yet. Reading the reviews you get, it's clear that just how heavy a sleeper somebody is makes a huge difference. Some people report the "teleportation" experience you want to deliver, some report a nightmare of bad sleep. Even with perfect soundproofing and suspension, there will always be this divide.

I am not quite clear on why disabled folks would want to ride your bus sleeping in their tied down chair, as opposed to any other bus. Of course, with solo sleepers, they should be inherently be highly accessible but if not, you can just dispatch one with the right functions. Or even an assistant can come with one if need be, who returns to the next job by robotaxi.

I am sure you have complied with safety regs, but I am curious as to how it works for a feet-first passenger. How do you crash test that situation? Are there belts? Are there airbags? Is the small berth sufficiently restraining?

As for efficiency, coaches like that tend to be about 6mpg, but there are some as high as 9. At 6 with 24 pax that's a very respectable 144 theoretical passenger miles per gallon. What is your planned load factor to get the real number? If you are running near 100% today, that's good but a sign of underprovisioning which won't persist. 80% would be a good goal, or a respectable 115 pmpg (diesel) which is about 95 pmpg (gasoline.) I think the solo car can get to better than that.

No worries Brad. Reframing the offer (for a future version of the service)-- If the Cabin stop was a 5-10 minute walk from your house, your sleeper berth was pre-assigned, you could board/get in in your cabin anytime between 9pm and midnight, and wake up/de-board anytime between 6 and 9am at a hotel where you could take a shower/change/eat breakfast, would that experience beat flying? Also curious how this equation for you would change for routes/destinations not well serviced by air (expensive flights and/or far away from major airport). Assume high quality active suspension and the sleeper berth being quiet (sound dampening + active noise isolation).

Feet-first is the standard berth configuration for entertainer coaches (e.g. and touring musicians say it generally feels more comfortable than the horizontal configuration (we interviewed one guy that spent 10+ years on tour, sleeping in tour buses on the road for >20,000 hours!)

Our target long-term occupancy is 80% (running higher than this now). One major advantage about having all your trips depart at 11 or midnight is that you can fluctuate supply based on demand. For example, if you sell 140 tickets on a Thursday night to LA, you can run 6 vehicles (5 of which are 100% full) and the 6th vehicle is 83%. This would be like if every plane from SFO to LAX left at 8am, the airlines could get to higher occupancy rates (majority of COGS are variable vs. fixed).

Curious what you estimate the fully-loaded cost per trip to be of the solo sleeper (toilet version and no toilet version). I'm in agreement with you that this provides the best passenger experience (if there are high quality rest-stops available along the route or the customer pays for convenience of on-board toilet) as it takes the user from origin to final destination in one transport vessel. I'm also a big fan of the work Chris Tacklind has done on Twill around passenger comfort/safety in solo vehicles--

It's tricky for the SF to LA run. This works on a shorter run, a run shorter than my sleep. If I can board any time, that's great -- if my berth is soundproofed from the other people making noise when they board or you keep really strict noise discipline on people once the first customer is berthed. But now there is the concern if it's more than 8 hours to arrival, since if I wake up underway, I don't want to lie there, but nor do I want to move to a seat for an hour of driving -- we have lost the teleporting. So if you can really do the ride in 6 hours this can work.

Yes, folks want to shower/change/eat when they arrive, or at least shower/change. It should be possible to find places to do that, though doing it for 24 people is harder. Hotels, obviously, but gyms might be a good option since they might not be busy at 7am.

Yes, if air is not competition, then you can have a win. But of course, if the route is popular, air should be there. I understand the advantages of feet first, but I am curious what safety testing has been done on it. Have people put crash test dummies in feet first crashes to figure out the best way to keep them safe?

For a solo sleeper, I see it being quite cheap to make (once the sensors are reasonable in price.) About 1/3rd the size (and materials) of a typical car today. To make it electric the battery is the most expensive component -- it would actually be cheaper to make it with an ICE right now but cost more to run. But at 100mpg, the fuel cost for a 350 mile run is under $15 vs $6 for electricity. $30 for repair, maintenance and insurance. So the rest is depreciation. Running 350 miles/day is about 110,000 miles/year. Unlike the 190,000 mile life of most cars, I think this sort of smooth running should make 3 years possible, 330K miles. At a $12K cost that's just 4 cents/mile or about $13/trip. If they can sell a Prius for $20K you can make this thing for $12K, probably much less. So a ticket price similar to yours is pretty profitable. In 3 years, $100/ticket and 312 trips/year is $93K. There are other costs of running a service, but in theory no staff (other than vehicle cleaning/maintenance and customer service.) And your NRE -- software developing, marketing etc. Maybe you could pull off an even lower price, I don't know if you have to.

Yes, Chris is a friend and I wrote about Twill years ago -- sadly they have not made the progress one would like.

Back in the 1970s, I had occasion to ride the Lake Shore Limited from New York to Chicago. This was a two section train--the other section came from Boston--and the two met up in Albany before following the New York Central route out to Chicago.

One of the offerings on this train--which was an overnighter--was what they called "SlumberCoach". This was an individual accommodation that was set up as a seat facing a small work table. Under the seat was a commode, and under the work table was a washbasin. Finally, if you folded down the back of the seat and closed the washbasin, you wound up with a narrow sleeping accommodation. It wasn't hugely comfortable, but it sure beat trying to sleep in a regular train seat, and at $10 over the price of a regular seat, it was a lot less expensive than a regular sleeping compartment.

I do not know how easy it would be to incorporate all these features into a car, but it really was a marvel of efficient and compact design.

Being able to convert is both a compromise and an extra cost. You need to be able to convert on a train or plane because nobody knows when they will be awake or sleep. That's even true in a shared sleeper bus. With robocars, if you want to sleep in a trip, you can just get out of your seated car and transfer to a bed car, and transfer again when done, within reason.

Which is not to say you can't find some economical ways to make things convertible. But I find they rarely make a great bed compared to a nice quality mattress. And there are all these other issues. When awake, you need windows and the ceiling needs to be higher. You need different passive safety. You want entertainment and may not be happy wearing earplugs.

Add new comment

Subscribe to Comments for "Sleeper cars"