On ultralight vehicles vs. large mass transit vehicles
One of the questions raised by the numbers which show that U.S. transit does not compete well on energy-efficiency was how transit can fare so poorly. Our intuition, as well as what we are taught, makes us feel that a shared vehicle must be more efficient than a private vehicle. And indeed a well-shared vehicle certainly is better than a solo driver in one of todays oversized cars and light trucks.
But this is a consequence of many factors, and surprisingly, shared transportation is not an inherent winner. Let's consider why.
We have tended to build our transit on large, heavy vehicles. This is necessary to have large capacities at rush hour, and to use fewer drivers. But a transit system must serve the public at all times if it is to be effectively. If you ride the transit, you need to know you can get back, and at other than rush hour, without a hugely long wait. The right answer would be to use big vehicles at rush hour and small ones in the off-peak hours, but no transit agency is willing to pay for multiple sets of vehicles. The right answer is to use half-size vehicles twice as often, but again, no agency wants to pay for this or to double the number of drivers. It's not a cost-effective use of capital or the operating budget, they judge.
The urban vehicle of the future, as I predict it, is a small, one-person vehicle which resembles a modern electric tricycle with fiberglass shell. It will be fancier than that, with nicer seat, better suspension and other amenities, but chances are it only has to weigh very little. Quite possibly it will weigh less than the passenger -- 100 to 200lbs.
Transit vehicles weigh a lot. A city bus comes in around 30,000 lbs. At its average load of 9 passengers, that's over 3,000lbs of bus per passenger. Even full-up with 60 people (standing room) it's 500lbs per passenger -- better than a modern car with its average of 1.5 people, but still much worse than the ultralight.
Light rail vehicles are even worse. They aren't light at all, weighing around 150,000lbs per car with 72 seats or over 200 people standing at full load. That means with everybody seated we're moving 2000lbs per person (worse than the average car) and fully loaded around 600lbs of vehicle per person. The DoE says the average train carries 23 people -- that's 6,500lbs of vehicle per person, more than even light trucks and SUVs.
As most of the weight in vehicles comes from metal, the energy cost of manufacture of the vehicles is closely related to the weight. For the ultralight single person vehicles, most of the weight will actually be batteries, and only the frame will be metal.
Stop and Start
All this weight is even more expensive because transit vehicles have to start and stop all the time. On electric lines, they can put some of that stopping energy back into the grid. Hybrid cars put some of their stopping energy back into batteries. Ordinary diesel buses don't do any of this, and even with the regeneration, the stopping and starting is highly wasteful -- but unavoidable for shared transit. Cars stop and start at traffic lights and stop signs, of course, but far less often. And the automated cars of the future, which will plan their trip and speed to glide along perfectly with the lights may not have to stop much at all -- and they'll do it with electric regeneration, slowly.
Moving weight is only one use of energy, of course. There is also air-drag and rolling resistance. The drag coefficient of trains (1.8) and buses (around 1) is much worse than that of cars (.26 for a Prius) or low-slung streamlined vehicles (less than .1). However, here, the sharing is a win. 9 people on a bus do better than 2 in a car or 1 in an ultralight, but not much better. 23 people on a train do even better. Long intercity trains do even better here. Indeed it is long intercity trains that are the most efficient vehicle we currently use, because at high speeds, air drag is the largest drain of energy, and long trains can have the least drag per person compared to short vehicles. Because loss to drag goes up with the square of velocity, jet aircraft, even though they are highly aerodynamic, lose out because they go so quickly. (Of course that speed is not without its upsides.)
Commuter trains are our next best bet. They don't go as fast, and they usually make only limited stops, and they usually run only at rush-hour so they have better load factors. Once they start having to run at all hours, they start to lose their gains.
Robotic single-passenger vehicles could also form virtual trains on the roads, following closely to gain much of the drag benefits of rail-based trains.
Rolling resistance also is not so bad for the large vehicles. Steel on steel, as found in rail, is 10x better than rubber on concrete, though light rail is not nearly as good due to both dirt and regular turns. However, rolling resistance also goes up with the weight, so a street car with much more weight per passenger loses the benefit it gets from being on rails.
The biggest loss of energy in transportation is engine loss -- energy wasted converting from the fuel source (gasoline, diesel, coal) to forward motion. Electric ultralights will do slightly worse than electric trains due to the use of batteries instead of overhead wires, but this difference is modest. Fuel burning buses and cars suffer more.
Next, transit vehicles run on transit lines. If your trip is not precisely along such lines, your trip is longer. Most transit systems are designed to move people to and from transit hubs, particularly a downtown. Yet the majority of trips in a typical city are not to and from downtown. Such trips require the use of multiple transit lines and often a lot more distance. In some cases, transit lines with private right-of-way may offer a shorter trip, but this is rare. The private vehicle trip is point to point.
However, it is true that if you want to reduce the number of private vehicles around by sharing them, there is some need for empty vehicle moves, which reduce the efficiency. However, if the vehicles are ultralights which weigh less than the passengers, the energy for these vehicle moves can be quite small.
It should be noted that studies show that transit users tend to take much shorter trips than car users. This fact has a complex explanation which is the subject of much debate. Some argue that the "transit lifestyle" encourages people to choose destinations for work, shopping and social life which are within a short transit trip, and that if everybody did it, we would all travel less and save energy. That cars have enabled us to live further from things and thus make us travel more. Others argue that there is a selection bias -- those who live close to downtown can take transit, those who don't can't. Both would agree that the slower speeds of transit (outside of dedicated ROW commuter trains which can outspeed cars) make it difficult to take long trips, but disagree on whether that's a feature or a problem.
Computer coordinated shared vehicles
Shared vehicles can be more efficient with computer coordinated carpooling, so that computers notice when multiple people want to take exactly the same route (full shared) or segments of the same route (Jitney) at almost exactly the same time. While this is more efficient, passengers will still pay a fair bit more for very low wait time compared to the added wait of shared or jitney service. At rush hour, finding people with common routes will be much easier. Solo and 2 person (inline) vehicles can be only 3' wide, however, which allows them to go 2 to a lane and possibly use cheap elevated right-of-way, while larger vehicles will be full-sized and constrained to existing roads.
Shared transit can be the most efficient system at rush hour, if cities are willing to devote vehicles and track/ROW to it, when the vehicles will not be used the rest of the day. Computerized operation could reduce the issue of having to hire drivers only for rush-hour. It's also possible that dedicated ROW at rush-hour could be released to computerized personal vehicles during the rest of the day, or even in between transit vehicles, which are usually spaced several minutes apart. In this case, the vehicles would need systems that can assure they clear the ROW in advance of any transit vehicle, which seems doable. Indeed it could be arranged that no vehicle would enter the ROW unless it had a travel plan which assured it would never impede a transit vehicle, with some slop included.
This analysis makes it a little more clear how 80 small single person vehicles can be more efficient than even a well packed bus or street car with 80 people on it. The 80 vehicles weigh less and stop less, though they have more drag and rolling resistance. The 80 vehicles cost less energy to manufacture, but if they are used only by their owners there will be a lot more of them. If they are shared (ie. they are taxis rather than privately owned) they will have a lower energy of manufacture, but a modest loss (worst case about 50%) for empty-vehicle moves.
Once you start comparing the individual vehicles to the average loads for transit vehicles, the energy efficiency victory of the private vehicles is much easier to establish. This is particularly true if the small vehicles are electric and make use of regenerative breaking, and clever navigation to avoid stopping and starting, as well as any need for sudden starts or stops. Based on DoE figures for transit ranging from 1,200 to 7,500 BTUs/passenger mile (3,500 average), and the figure of 300 BTUs/passenger mile for existing simple ultralight designs, the personal vehicles are 5x to 10x greener.
Even "light" vehicles (such as the Tesla, Tango and many others) which meet highway safety standards still do quite well at 1,200 to 1,500 btus/mile and 800 to 1,200 btus/passenger-mile when shared, if they can be shared. They still beat U.S. transit, though not Asian transit.