Robocars: Deliverbots -- computer driven trucks


For part seven of my series on Robocars, I now consider the adjunct technology I am calling Deliverbots -- namely robot driven trucks and delivery vehicles, with no people inside. These turn out to have special consequences of their own. Read:



This deliverbot article is captivating! Once you start thinking about it, the ramifications really become mind-blowing.


Great site. I'm only part way through it, so far.

Deliverbots will give us the "physical internet", which I believe will change everything. As well as what you speak of, they can also function to work as a go-anywhere production-line so as to change a great deal of our industrial operations too.

Just some comments:

PRT: I agree with most of your thoughts on PRT. I think the ULTra system is the most promising PRT system, because it can expand out onto the roads and evolve into a "roborcar" system in its own right. And the advantages of that, like you've said, are overwhelming.

ULTra also has the potential to function as a 'horizontal elevator' for new residential developments. It can't take trucks, but does it need to for new developments? We don't need to take a truck on an elevator that supports the apartments of a high-riser, we just break down the parcels from the truck and send each small individual item to each apartment. PRT can do the same thing for 'the last mile' in a residential development. The included link is my own idea here.

Indeed, the demand for high-capacity trucking could be dramatically reduced with robocars because you don't have to worry about the cost of the driver overhead, so you can move in the direction of individual vehicles for individual items. Trucking could mostly be reduced to intercity operations only. In many ways small but large numbers of robocars could be more efficient than trucks due to the lack of intermediate stops. Certainly more energy-efficient.

Rail versus rubber:

Just thought I would point out that rubber can be made to be much more energy-efficient in a more specialised system. You can make wheels that move closer to the direction of push-bike wheels, which reduce rolling losses and to the point where those losses are as good as irrelevant. (this means relatively large diameter wheels, with a thin rubber ring. And also maybe high pressure tyres and low rolling-loss rubber). You can probably get away with this if the cars are operating in more controlled conditions, and in a more controlled way i.e. they don't have to deal with the extremes of what a traditional cars has to deal with.

You might like to check out my own closely related ideas relating to this:

All the best,

Andrew Atkin

Hi --

It's interesting to read both your musings, based on two different systems (robocars and ULTra); since I did the same with yet another. I'm a big fan of Doug Malewicki's SkyTran proposal (currently in joint development at NASA Ames: While living in Paris a few years ago, I created the site dedicated to exploring social ramifications of that system. All three of us came up with quite a few similar ideas, including the revolutionary impacts of automatic delivery.

I have a background in controls; so I did a lot of thinking about how cheap and fast automated goods transport will change production and distribution. Despite some differences (SkyTran and ULTra avoid road congestion; SkyTran is faster but limited to a grid every few blocks), the basic effect is the same; so I invite you both to compare your concepts with those I and others put there:
1. Extreme Just In Time: not just giant firms making a few defined products such as cars, but most manufacturing will move to lean production networks that span entire metropolitan areas and produce an incredible variety of products on demand in a few hours. Doug loved "Little Shop of Wonders": using VR goggles and a small selection of physical coats to choose a garment customized from the thread up; then manufactured and delivered in a few hours.
2. The resulting Extreme Customization and Extreme Outsourcing will re-shape urban areas. You choose products in an urban store, but they are actually delivered from a low-rent warehouse or manufactured to order.
3. Dense cities will become centers of manufacturing again (goods can traverse New York much more quickly than LA; so manufacturing can be much more deeply customized).
4. Networks will eventually extend (at a wide grid density) even to rural areas. Same-day shopping will transform rural life, and it will be even easier to live in a distant suburb and work downtown (and on the way to and from). However, I speculate that cities will be unleashed even more and become the Best Place To Be (do the titles give you a clue that my original wiki was Qwiki ;-) . By facilitating a sophisticated urban lifestyle that will attract the rich and also creating large numbers of jobs for the masses, such transportation may lead to substantial re-urbanization and improved Class Welfare.
5. As Andrew proposes for developments, Skytran/ULTra/robocar fleet Equals Elevator allows virtual physical offices and stores spread over miles.
6. Same-day custom manufacturing and the ability to have a face-to-face conversation in minutes are advantages no distant low-wage country can replicate: Distance Revived.

Regardless which mix of systems we end up with (as Andrew says, ULTra vehicles might leave their guideways; robocar taxis and delivery would complement SkyTran), I think we are working with powerful ideas here. Enjoy!

Dr. Howard Goodell
Boston, MA

I do feel that on-the-street delivery robots are much more dramatic than a track-based system, because they go everywhere, while the track system has to start small and slowly get bigger but never goes everywhere. (In both senses. It never goes to people's houses, and it never goes to many neighbourhoods either.)

But the track system could go to malls and factories, offices and apartment complexes.

I'm not so sure about #3. Yes, if the factories are densely packed, goods can move quickly. (Manhattan still has the remnants of a pneumatic tube system that was even faster for very small packets.) But is there a big difference between 2 minutes and 20 minutes? If you're literally sitting waiting for it to come, unable to do anything else, then of course it's a difference but I don't think there's a big advantage to designing manufacturing processes to be that just-in-time. JIT is great but I think there are diminishing returns. Especially if you can order a few extras when you are getting low and even return ones you don't use if you are really short of space.

For #4, as I said, deliverbots extend to the rural areas the day they can be safely deployed.

On #6, I fear videoconferencing is going to improve to be quite suitable for business conversations, if not for personal ones. But this one could go either way.

I and others have puzzled over this question. Do the robocars make you want to live even closer or even further away? Nobody yet knows enough to predict this future.

Absolutely a mix of computer driven and human controlled driving can work. The issues, from a technical view, are not insurmountable. We have autopilots that can land commercial jets. This was no less difficult a task. The issue of safety is top on the list, and those dealing with the problem will have to ensure public safety. They of course know this. The next issue is economic. Is there a monetary benefit to having computer controlled vehicles? If the answer is yes, we will eventually see the roads covered with driverless cheap cars. We solve problems mostly based on incentive.

I have left the above comment in, for those who read the comment threads, because it was a spam. The word "cheap cars" in bold had been entered as a link to a car selling site. What's amazing is that they're paying people now to try to think up meaningful comments and then insert spam in them. This comment seems original, not a copy of another one on the thread etc. but perhaps it comes from elsewhere.

And the spammer smart enough to write this ignored the warning that says that links get nofollow and thus will not improve search engine rank as this was intended to do.

Back in 2011, I displayed a mock-up of my idea ( "my idea", which, as we see here, tend to be unoriginal in a world of seven billion people, some with plenty of time for pondering the future) for a "Driverless Errand Car (DEC)" at Maker Faire in San Mateo California. I fell short of getting the thing to drive around, even in a circle, as a demo, but people listened to my ideas regarding a car with no passengers. What makes a DEC so attractive, is its 24-hour availability. It's like adding another self. Rather than attempting to run around town after work, getting prescriptions, dry cleaning, milk, cat food, pizza, and the like, a DEC vehicle can proceed very slowly at 3AM, make stops, and return, all while you sleep. Unlike some of the mockups pictured here, my vision is for a vehicle that can climb stairs, get over obstacles, and doesn't need a 'pool table surface' to operate. My base criterion is that, being sick in bed with the flu, the DEC can leave the house, pick up suitable food and drink, and return to my bedside so I don't have to get up if I don't feel like it.

Large items? A couch, new desk, or refrigerator? 2x4s? sheets of plywood? I envision that robots will become cheap enough that we will own five or six, and that between two and four transportation robots will work together, with a specialized carriage suspended between them, that they will transport almost anything. Speed won't need to be a factor, with 1AM to 6AM pick up and delivery. And, in a further refinement of Uber, Lyft, etc., when you are not using your transportation robots, they can be assigned to other people to transport goods in the return direction (i.e., no "dead-heading"). Considering that our automobiles spend 93% of their functioning life, sitting parked, the energy requirements now demanded in the USA for 200 million vehicles extant, will probably drop by 70%. The automobile, with the unitized driver, powerplant, passenger compartment, and ownership (insurance, maintenance, sunk costs in the thousands) will disappear.

Add new comment