You can get something better: neato. It uses LIDAR.

If you have solar panels capture the energy, they simply suck the energy up, store it, and when it returns as heat in the friction of the objects it moves, the lights it powers, etc. Without the solar panels, the light would just be heat. So it is free.

Which university?

The main issue with public transportation energy efficiency is occupancy. Trains and buses in a "walk-on" schedule are not guaranteed to be fully occupied. Airlines and highspeed rail operators manage demand so that trains and planes are often up to 70% occupied. However, walk-on buses and trains often have under 20% occupancy. What this means is that cars do a lot better except in rare, properly-managed, high-density scenarios. Here in the USA, outside of a few urban areas, such San Francisco and New York, there simply are not the densities needed for successful public transport operation. In these cases, the automobile (and motorcycle) are actually more efficient because of occupancy.

Case in point. A 55 passenger advanced hybrid passenger bus gets 5.5 MPG city. With all seats full, it gets 300 pMPG, which is really good. But, this almost never happens. Why? Because the average number of people on a bus in the USA is 9 (UK is 10). What that means is that the bus gets 55 MPG. A Prius with one driver gets 50 MPG, which is similar. A Prius with 5 people packed in gets 250 pMPG, similar to the bus (I'm sure the Prius would do better if it was a diesel hybrid).

In a suburb, or in a rural area (like most of the US!), most of the trains would be empty. Rail also has very high costs (much higher than building a road and running the cars) associated with it if it cannot be fully utilized. Should people live in suburbs in rural areas or should they give them up and live in a city because of transportation efficiency? That's a different question. What we do know is that cars are the best way to deal with suburbs and rural areas, and buses are a necessity for those who cannot afford to own cars or cannot drive cars because of disability. In the future, if robotic cars become a reality, we can have robo-taxis instead of buses for the disabled.

Actually, the train is already maxed out in efficiency. People don't understand how efficient electric vehicles are. Well designed ones, such as that Japanese train and Tesla Roadster are 90% efficient or more. There's not much room for improvement in either case. Many people are building their homemade electric cars on A/C induction motors which are used for industrial equipment and trains - one electric motor that powers on axle of a train can easily power a car. It might actually be more efficient (but not by much) because oversized motors are used at lower current, which reduces cooling and gives the car more HP, which is always good.

This really demonstrates the fact that switching to alternative energy sources (switching from gasoline to electric) can have better effects than conserving energy by switching from one mode of transit to another.

BTW, in the USA, there is an average 1.54 passengers per car, so to make it fair to the car, the car uses 100-170 watt-hours/passenger-mile.

Yeah. I was trying to type 250 watt*hours/mile, not 25. Ironically, 25 is right for a moped. Leaf has a 24 kWh battery pack and goes 100 miles/charge, so I think were in the ballpark.

Aluminium also contains 2x as much energy as gasoline by volume.

That number, posted above, is inflated. It is really 25 watt*hours/mile.

That figure is inflated by a factor of 5 - it's really 16 kWh/100 km or $1.6/km. In the USA, needs to tax gas at $0.60 a gallon to pay for roads (it currently taxes it at $0.40~ish).

That number is wrong, because it calculates the energy in the gas that goes into gasoline cars. 80 kWh/100 km is 1,287 watt*hours/mile, five times higher than the average EV highway energy rate of 250 watt*hours/mile - electric cars are 5 times more efficient at using energy than gas cars. At low speed (city driving), a EV consumes around 150-160 watt*hours/mile, similar to the Japanese rail system, which gets 150 watt*hours/passenger-mile.

You get all that (except maybe tech support) for $0 with Ubuntu.

It also cost $100 or so a year to be in the app store.

You don't need to. I'm gonna anger a lot of people by saying this, but physical fitness is not really important. In a world where one can buy a $600 dollar part that can convert bioenergy into mechanical power at a 10X the power to weight ratio of a human being, it's not really all that important.

Oh, and mindless junk is important. Video games are the reason my first job was coding, not washing dishes. I was allowed unlimited game time assuming I finished all my school work. Guess what? I got bored of the games. So I wanted to write my own. I did. Then I discovered my coding skills worked for robotics and web dev. Then I found out that I could write applications for doing "business" stuff. So, one of the reasons I am where I am today is because of mindless junk.

Really, I can't wait for the first DIY experimenters to hit the highways in their homemade robocars. Slogan: DIY Robocars - making safety dangerous - your car, your code, our road.

Actually, users won't have to. People think that robocars = no human drivers, but I don't think this is really true. All they need to do is have the same algorithms, but give the human drivers say 1 foot of play back and forth in the lane and say 2 seconds ahead and behind. Then people will still feel like they're in control, driving will still be fun for those of us who enjoy it, and those of us who would get sick to our stomachs in a robocar would feel better.

Under normal driving conditions, a human is not really better than a robot. It's the extreme conditions, where the driver is not paying attention to the road, or a sudden even, or ice and snow or a mechanical problem that gets you. If you build a few avoidance systems, you could make driving a ton safer without going through all that work of building a full robocar where you get in, click on a map, and sit back. You don't need to go through and figure out how to drive every weird unmarked intersection in the country, "just" look at some laser scans and hit the brakes if something gets in the way. You also get superhuman vision, like infared cameras, laser arrays, sonar, DHS xrays, as well as road information such as traffic cameras and CCTV's.

You have the following:
1. Fully human-controlled (1980's and before)
2. Computer controlled emissions equipment (1990's)
3. Computer controlled stability and traction (2000's)
4. Computer assisted accident avoidance (2010's)
5. Optional full computer control? (2020's)

Notice that there is no mandatory full computer control. That's because I don't think you really need it. Sure, some economy cars will come without steering wheels or stuff, but many luxury and performance vehicles will. I think what will happen is that you average sports vehicle, be it and offroader or a sports car or a road-going rally car will have three settings: no assistance, computer assistance, and completely autonomous. Eventually, "no assistance" will not be a good idea on public roads. Not like it is going to negatively impact the driving experience anyway.

Now, many people discuss scenarios where "road trains" and other modes of driving that would not be possible with human drivers with no spacing between the cars. I don't see that as a viable scenario. While many accidents are caused by people's mistakes, some are caused, or at least helped along by mechanical failure, such as tire blow out or brake problems. These automated cars may be better at reacting to the problems, but I don't want to be in a road train inches in front of a semi when its tire calls it quits.

If robocar technology can save 40,000 lives a year, why should we care about letting people drive them? Because it will allow the systems to make a significant impact much earlier, and it will cause a who group of people (people who like driving) who would be otherwise opposed to robocars to be supporters of robocars. They're actually a surprisingly useful group of people, being car mechanics (who will fix the things and install assistance systems on pre-built cars), race car drivers (who will promote the things to the general public), and not to mention people who work for automakers who build the cars in the first place.