Brad Templeton Home

Main Page

Brad Ideas
(My Blog)

Robocar Blog

My Forbes Articles

(Book from 2008)

The Case



Car Design






Future Transit













Urban Planning



End of Transit




Sidebars: Charging

Speed Limits

NHTSA Levels

Valley of Danger


Google Cars

New design factors for Robot Cars

New design factors for Robot Cars

Read/Leave Comments

(This is a special article summarizing how the design of cars can change due to robocars. Consider the rest of the site for a more top level introduction to the big concepts or the blog for the latest items.)

Many of the big changes that will come about from robocars will come from how they free car designers from the constraints of human-driven cars which are the owner's sole, or almost-sole vehicle.

Much of this depends on this yet-untested idea:

If one can hire a cheap specialized "robotaxi" (or whistlecar) on demand when one has a special automotive need, car users can elect to purchase a vehicle only for their most common needs, rather than trying to meet almost all of them -- or to not purchase at all.

For example, for many, most trips are short, have only one passenger and do not require significant cargo room. Almost nobody purchases a vehicle good only for that purpose, because they want to cover the occasional needs for long trips, taking extra people, carrying cargo, towing or going off-road.

Some of these changes would also apply to what I'm calling "whistlecars" -- owned or rented cars that deliver themselves to you when you summon them, but which you still drive.

Let's consider some of these changes, and the constraints they remove from vehicle designers.

Range is much less important

Today, car buyers demand long range. Electric cars are mostly of modest range, and worse, after they reach their range, must find a rare charging station and sit at it for many hours. This is probably the greatest impediment to their sale.

A robocar need only meet the range needs of most of its owner's usual trips, since a longer range car can be called for when needed. It is also possible to combine vehicles, so a short-range robocar might take the passenger to a longer-range robocar, or to an intercity train.

Battery problems are considerably reduced

With no need for long range, electric vehicle design overcomes its largest challenge -- the batteries. Decent range demands lots of batteries, and lots of batteries means lots of cost and weight, with significant expense and recycling trouble required to get lower weight batteries.

Short range electric cars can have minimal batteries, and can thus get away with cheaper and heavier battery technologies. Other short range clean power technologies, such as compressed air, also become practical.

The power train itself becomes relevant only to the fleet manager

Taxi customers have no reason to care about what the vehicle's power train is, as long as the vehicle can get them from A to B at the desired speed. Customers might care about how green it is, but generally it is only the fleet manager who will examine the various issues around choice of power train and fuel.

Refueling is not usually done while humans travel

A robocar or whistlecar will refuel/recharge itself after dropping of its passengers. Fueling/charging stations need not be conveniently located on major travel routes. Vehicles can travel a modest distance (vacant) to reach them. (As they do this, energy is wasted, but human time is not.) This allows slower and more expensive refueling systems, including pressurized gas refueling (hydrogen, LNG or even compressed air.)

It also allows -- at an energy cost -- early adoption of new refueling technologies without having to put stations everywhere. As long as trips to stations do not require a significant or expensive fraction of the vehicle's range, stations can be rare and still work. (The lack of suitable refueling infrastructure is often cited as the main barrier to several alternative fuel technologies, such as hydrogen.)

Because refueling/charging will be done just prior to trips, and the energy will be expended immediately after, some alternate energy storage methods may become practical, such as leakier hydrogen storage, supercapacitors, flywheels (even lossy ones) or liquid nitrogen.

Electric vehicles can do slow recharge, at a cost of increased downtime. Fancier high-current charging stations can offer fast recharge. Also possible, and more suitable for robotaxis, is battery exchange. Battery exchange, with a standardized battery module, can be human guided or done by robots -- we are, after all, presuming many general advances in robotics on the way to robocars.

Electric recharge (even of removable battery modules) can also be moved into the night, where power is much cheaper. In some cases, such as coal and nuclear plants, electricity demand at night is a must that can be well-met here. Alternately if most power became solar, daytime charging would be done.

(Obviously there are limits to this, and in an all-electric world, battery packs will have to be charged at times that balance load. This calls for cheaper batteries, so you can pick and choose.)

Single passenger vehicles will be much more common

Most trips will have just one passenger, so single person form factors become open to car designers. Single-width cars can fit far more easily on roads and reduce congestion. Dual person cars can be designed both side-by-side and inline, including inline face-to-face, thus still taking just half or 2/3rds of a lane.

Because robocars can negotiate with one another, 3 cars, each 2/3 of a lane, could work together to fit in 2 lanes.

Reverse and face to face seating

In general, face-to-face seating will be a popular choice for multi-person cars of the 2 and 4 person variety. Single cars may also offer backwards seating. While this will be reviled by many, it offers two advantages. First, in any front impact accident, sitting backwards can greatly reduce the risk of injury as long as pre-impact braking has pushed the person's head and body against the seat and headrest. A seat-belt is possibly not even needed here. Secondly, the most aerodynamic shape is the teardrop shape which is fat at the front and a point at the back, and in a very small car this may fit a backwards facing person better.

The steering wheel vanishes

Over time, the need for an obtrusive steering wheel goes away. For times when a human driver must take control, a joystick may be adequate if the computer is working well. A pop out handlebar which unfolds and locks could even provide mechanically linked steering, but take no space during normal operation. A brake pedal is of value for emergency operations, but the throttle could also be a hand control if meant for only emergency or occasional use.

Windshield requirements are different

While a nice windshield may be good for visibility for forward-facing passengers, there is no need to have a large unobstructed view for safety. The windshield can be reinforced with bars, for example, allowing it to be much stronger in the case of impacts, notably impacts with animals. Other than for passenger comfort, the windshield barely has to be there at all.

Headlights are also not necessary, at least with current sensors. However, for peace of mind, some passengers will still want to be able to see where the vehicle is going.

Cargo space is not necessary in all vehicles

Those wishing to move more than personal cargo can call for a robotruck/deliverbot with the capacity they need. If they already have their personal robocar on hand, the truck can follow them to their destination. Otherwise they can call for a robotruck with passenger compartments, or a car with cargo capacity.

Acceleration is not a big requirement

While electric vehicles all have good acceleration, an ideal robocar trip is perfectly timed with traffic lights and other traffic so it does not stop and start regularly. We like this because it's more efficient, but it also means that acceleration is rare, and need not be that zippy. Indeed, for comfort, you may prefer it slow.

This may allow transmissions to be designed differently, to be cheaper and more efficient -- or even non-existent.

Sport driving vehicles will continue to have good acceleration, of course. Whistlecars would probably want this acceleration too.

Today, the price of a car is often strongly linked to its acceleration. This may change.

Speed may not be that important

Vehicles meant just for urban trips need not even be capable of highway speeds. Vehicles only for long trips need not be fully efficient at slower speeds.

If passengers find a pleasant working/talking/reading/viewing environment in the vehicle, trip time may become less important than comfort. If the passenger can work efficiently while in the vehicle, they might accept a longer trip to be cheaper or have fewer stops. Such vehicles do not need to be fast. Save that for rental sporstcars.

Different needs for speed and acceleration may allow entirely different engine and transmission designs.

Cars may be much lighter

All of the above factors allow the car to be much lighter than today's designs. In addition, once the risk of crashes is greatly reduced, more weight currently devoted to safety systems can be reduced.

Smaller battery requirements, motors, inability to go on highways and being single seat all point to a much lighter -- and thus more energy efficient -- vehicle. Indeed, a single person light urban vehicle can be built today that uses 1/10th the energy per passenger mile of today's cars and transit systems, and 1/3rd that of Japanese trains.

Suspensions can be super-soft

Human driven vehicles (HDVs) want suspensions which transmit the feel of the road. A robocar should have a suspension that eliminates bumps & vibrations from the road as much as it can. It may even have a computer-controlled suspension, using shocks with electromagnets or ferofluids, combined with a scanner which examines the road surface ahead of the wheels, and adapts in real time to eliminate the effects of potholes or other problems.

One can also conceive of having more than 4 wheels, so that any one wheel can be decoupled from the body as it goes over a small bump or hole.

Vehicles may also mount the passenger compartment on arms so it can be pitched in turns or tilted on acceleration and braking to provide minimal distraction to the passengers.

In time, safety concerns change considerably

Robocars will be highly unlikely to get in accidents. Once this is proven, their designs can be made lighter, and many safety features may not be as important. (This is not to say that accidents become impossible, but they may become so rare as to change the economic trade-offs of these safety features.)

As long as human drivers share the road, safety belts will presumably be required, as a robocar may need to brake or swerve suddenly to react to a sudden stop by a vehicle ahead, road obstacle or pedestrian on the road. Airbags, however, may be able to suffice if such events become rare, as they will on highway lanes where pedestrians, and eventually human-driven vehicles (HDVs) are forbidden.

Today we allow passengers on trains and buses to not wear seat-belts, and to get up and walk around during travel. As robocars attain the safety record of a bus, this can be allowed in them.

Also possible if human driving is never intended are vehicles without forward view windshields (or with partially occluded forward views.)

The in-car environment changes considerably

The in-car environment will become more of a work and entertainment space than just a travel space. Passengers will expect things like a screen, a keyboard, and a desk. Passengers may wish to face one another (though not all are comfortable riding backwards.)

Quiet will be a very important consideration, though passengers will be allowed to wear headphones if desired, unlike drivers today.

Aside from working, people will be doing new things in cars -- all the things that passengers do, and a few more. They will do these to a larger degree because while in the past a passenger might eat while the driver drives, that's a bit anti-social. However, dining together (at a table) will probably be common.

Sleeper cars

The smooth ride (especially on the highway) of a robocar may generate demand for cars for night-travel, while the passengers sleep. Such vehicles might aim to make a trip last 8 hours rather than make the fastest possible trip, and as such would be much more energy efficient for such trips.

(This also requires a very low crash rate, as seat belts don't work as well on flat beds.)

While a sleepercar could be a whistlecar, it would only work with a series of drivers who could take shifts.

And yes, people will do all the things people do in beds today. Eventually we may get negative fatalities from cars.


Reversing the green trend of the short-range electric vehicle, there will be demand among the wealthier for RV type vehicles. Robo-RVs may be larger than today's RVs due to the computer's ability to safely move a larger vehicle, or an articulated one.

It also becomes possible to have a "multi-unit" RV, where several vehicles travel together and then dock or simply park closely together at a campsite. It may also be possible to rent additional modules in local areas and only move a smaller main module along with the people.

Taken to extremes, some may create an entire "mobile home" consisting of many towable units which can move to a location and dock. Because of the energy cost of this, and the fact that they can be sent ahead by slower means, such mobile home units might move themselves to conventional railways for the longest part of their journey. Railways are the most energy efficient land transport. (Sea transport is even more efficient through routes are longer.)

Of course, in many cases it would make sense to instead have permanent buildings with the desired extra space and facilities, and have the robo-RV with the user's personal gear dock to that, but such motels would not be found everywhere, of course. It becomes a question of whether the added cost of hauling your own facilities is worth the added convenience and customization, compared to the cost of renting local robo-RV-modules or motel space.

Parking is not a problem for the humans (or society.)

Robocars will park themselves (if they are not hiring out for more work) after dropping off their passenger. They can park some distance from their owner, at a cost of energy and slight delay when the owner requests the vehicle come to her. They can park far more densely in existing parking areas, and half-width cars can park yet more densely. If there is a good estimate for when they will be needed (commute times, fixed length events) they can park, if need be, remotely for a while, and then park more closely around the estimated time of need.

Parking structures could also be built for robocars that have many "half height" floors since humans will not walk frequently on the floors. However, it is debatable if new parking will actually be needed since robocars will park so efficiently, and often be hired out and not parking at all in dense areas. The more cars are shared, the fewer cars there are total in the city.

Robocars/Whistlecars can store themselves in many places HDVs cannot, such as at the entrances to driveways on city streets, since they can always accept a request to temporarily unblock the entrance. They can also make dynamic use of street-sides, if necessary taking up every lane but one during low traffic periods, but clearing out when traffic increases. It is not a problem to have robocars "double park" and even "triple park" as they will clear out on request.

Individuals need not have garages in their homes, even if they own robocars. Their robocars can find dense parking somewhere near the home, and come to the door on very short notice. This in turn changes many of the rules of how we design buildings around cars, offering more useful street frontage in the homes.

Zero emission vehicles may also be allowed to enter buildings to drop off people or cargo.

You can read full notes on parking for more details.

Many car owners may rent out their cars

A new market will probably develop for people who wish to own a robocar, but wish to rent it out when they are not using it, either to be greener and help pay for it; perhaps even profit from it.

As such, these vehicles would be designed to facilitate such rental, including cameras which can photograph the interior before and after (but not during) any rental to identify who does any damage.

Cars will of course be able to take themselves to car cleaning stations as appropriate, when not in use.

Also desirable will be a special lock-box for the owner's personal gear that they like to have with them when they travel. It should be easy (with the keys) to move this lock-box to another vehicle, or to have a robot extract it from the owner's vehicle and get it delivered to where the owner is if they need something from it. The boxes would come in a few standard sizes to make them easy to move from car to car, but it might also be possible to make one almost as big as a regular trunk that moves only among a limited set of cars, or involves swap-out among a set of users of one specific car.

Car clubs

Buyers may wish to form car clubs, where they share a pool of vehicles but rarely or never rent them out to strangers. Such clubs would allow members to express themselves through their vehicles without the cost of a wholly private vehicle. In effect this would be like buying a "time-share" of a car, except it comes to you when you call it.

Car-clubs might form ad-hoc around a single vehicle. For example, one might join a club with no other purpose than to share one or more of a certain hot car, and belong to another club to share a different car.

In all cases, members of car clubs could still rely on hiring outside vehicles if 2 members want the same vehicle at the same time.

Cars would be designed expressly for car clubs, with ways to do automated customizations for the particular occupant.

Age vs. Miles

Vehicles that are shared (as taxis or otherwise) will see far more use in a day than typical cars do today. As such, their components will wear out much faster, and generally will wear out because of miles, not years, though interiors will wear out based on hours of occupancy. Components should be redesigned for maximum active usage without concern for aging.

More cars will exist that perform only very specific duties. There will be SUVs that do nothing but rugged driving. There will be city cars that never go on the highway. They should be designed for these use patterns.

Cost concerns

The ability to share vehicles in this manner may alter cost equations in other ways. While full-time taxi companies will choose vehicles that give them the best return, private buyers may be willing to buy more expensive cars due to the return they get on renting them or sharing them.

You can read a week of stories of a robocar future or jump ahead to consider the downsides of robocars.

Read and leave comments on this page.