Automatic Delivery Systems
Up to: What future do we want?
An automatic delivery system is a system for the
transmission of material objects between homes, stores, offices etc.
with as much as possible of the convenience of the telephone system
used for the transmission of information. First we shall discuss
how such a system might look to a user. Then we shall discuss the
advantages of such as system and what it might be worth. Finally, we
shall propose some ways of implementing the system and try to
estimate how much it might cost.
(Except for specifically designated 1995-96 notes, this essay was written
in the 1970s.)
How it looks to the user.
- An apartment or home or a department of a store or an
office has a port into which an object may be put (perhaps in a
suitable container) and a push button system for dialing
destinations. After a while the object arrives at its destination.
To relieve nervousness that I might be wishful thinking about matter
transmission by telekinesis etc., let me say that I have in mind a
mechanical system that will transport things through tunnels under
the streets.
- I can send an object to a friend, and it will arrive at
his port in a time comparable to the time required to deliver it by
car.
- A store can send me something bought by telephone. They
can send it right away, because they don't have to wait to fill a
truck bound for my part of town.
- If we also have home computer terminals and electronic
transfer of money (both of which are easier to realize than the
delivery system), and if the seller is an automated warehouse rather
than a store, then I can order an object at any time of the day or
night and have it delivered immediately.
- I can send and receive mail by the system. Long
distance transmission gets switched to other modes of transportation
at suitable places.
- I can get rid of trash by sending it to the trash place.
- I will need to own fewer things, because I will be able
to borrow or rent them with less transportation overhead. A group
of people are more likely to own something in common if they can
readily pass it back and forth.
- Home delivery of cooked meals and return of dirty dishes
will be much more feasible than it is today.
The automatic delivery system will be of special benefit to
children, the handicapped, and old people since it will make them
more independent; it will extend the age and decrepitude at which
people can live independently. I appreciate the point particularly,
since I expect to be quite old before the system comes into use.
The idea of an automatic delivery system is quite old in
science fiction, and its advantages are apparent. It was certainly
infeasible 60 years ago when we got our last batch of public
utilities. Let us consider whether it is feasible today or will be
in the near future.
How it will work.
First of all, we shall try to devise a system that will work
with present buildings. Something that requires new buildings will
be quite hard to implement.
Therefore, imagine the following: The ports are mounted in
outside walls or in windows like air conditioners are today. This
requires minimal modification of buildings. The carriers do most of
their travelling under the streets on continuous belts or suspended
from cables, but are independently powered by rechargable batteries
between the under street system and the ports. They move from the
under street tunnel to the building through feeder tunnels and climb
the outsides of the building to and from the ports. There are
several ways this can be done having different divisions of the
investment between the building modification and the carrier. If we
want to put the investment in the carrier, then the building is
equipped only with ``handholds'', and the carrier climbs the building
with two suitable arms. If we are willing to mount cables or rails
on the building, the carrier can be simpler. On the whole, it seems
to me that the ``handhold''system is better, because it is more
routinely adaptable to a variety of buildings and it will make the
minimum change in the appearance of the building. Any version of
the system requires an elaborate system for switching the carriers
at the right time. This can be done by a a computer in the carrier
which communicates with the central computer controlling the system.
An important characteristic of the system affecting its
utility and cost is the size of object that can be transmitted. A
reasonable size carrier might be rectangular with dimensions
16"x16"x48" having an internal space 12"x12"x36". The carrier would
change orientation as it traveled so the contents would sometimes be
upside down and accelerations of say 3g might have to be tolerated.
It would be desirable to design the system as a whole to accomodate
a range of sizes of carrier and so that parts of the system could be
upgraded to allow larger sizes. One would probably want a lot of
quite small carriers for mail and single small items, but the size
mentioned above probably should be provided for in any case.
There are some safety considerations. First, there need to
be guard rails to keep people and animals away from where the
carriers come to the surface. Second, the carrier needs a sensor to
detect that the next handhold is available and to detect excessive
resistance to its motion. Ice, fallen tree branches, and damage to
buildings are the most likely causes of blockage. In such a case,
the carrier should stop and call for help. Perhaps it will be
desirable for the carrier to have a TV camera so that a human in the
control station can see what has happened and decide what to do. A
smart program in the central computer may be able to decide some
cases without human help.
The traffic capacity of the system will be a cause for
concern. If everyone orders his dinner through the system at once,
there may be bottlenecks. Capacity can be increased by providing
parallel paths under streets and by having small carriers ride
bigger ones until they have to branch off.
The system is obviously most cheaply constructed for a city
full of new apartments, but it looks feasible even for present
suburban areas, though at greater expense.
How can we estimate the expense?
A few man years of mechanical, civil, and electronic
engineering could produce an estimate accurate within a factor of
two with an uncertainty of a few years in how long it would take to
get a system working and an factor of five estimate of the
development costs. Clearly it won't be cheap, but I think we will
be able to afford it in the next ten to twenty years. Here are some
very rough estimates.
- The carrier is perhaps the easiest to estimate, because
it can be compared to a car. It is much smaller than a car, and it
spends most of its time riding. However, it will need a more
complicated control system than a car. Therefore we estimate its
cost at $500 taking into account expected large reductions in the
cost of electronics.
- The port. Adding a port to an old building, we will
guess at $500 for an old building and $200 if put into a new
building. The cost of the handholds will depend on the height and
shape of the building but shouldn't be more than $200 per port. The
cost of the feeder from the street to the building, we estimate at
$1000 to $3000 depending on the need to tear up sidewalks and
streets.
- The communications cost is estimated at $100 per port
assuming it piggybacks on the telephone system.
- The central computer cost for a city at present prices
might be $10,000,000, but this will go down. [1995 note: It went down.
If only the political problems would go down.]
- The biggest cost is likely to be the under street
system. It includes a tunnel whose size determines the possibilities
for expansion in size of object and volume of flow. It has to
provide for two way traffic and to carry this traffic suspended from
a cable, on a moving belt, on cars on rails, or simply to provide a
right of way for the carriers if these are independently mobile. We
shall suppose that the carriers contain the information and computer
facilities for deciding when to switch paths.
The cost ought to be considerably less than the cost of the
streets themselves in new districts since the carriers will be more
efficiently loaded than the cars that now provide delivery services.
Suppose we guess $1,000,000 per mile, but a civil engineering cost
study could make this more precise.
- Stores and warehouses will require more elaborate loading
systems than homes or offices. Presumably, they will be designed to
summon carriers and load them automatically. Such systems will cost
from tens to hundreds of thousands of dollars and perhaps millions
for institutions like the post office, Sears Roebuck, a railroad or
air terminal.
Well, suppose the technology is ready in five years to make
an economical automatic delivery system. (In five years, we expect
electronic technology to make the communication and computation
cheap, but we don't expect a cost breakthrough in construction
technology by then.)
How can an automatic delivery system come into
existence?
The system will be a public utility and a natural monopoly
like the telephone system, electric power, gas, water, sewage, and
roads. It could come into existence either by a sequence of
engineering studies by the government and establishment of a
government operated system or it could come into existence as a
regulated utility operated by private enterprise. No important
public utilities have been established recently, so it will be a
new political issue. As I see it, it is more important to society
that the system be built than who builds it. However, it seems
more likely to be built by private enterprise, because if it is to
be built by the public, there must be overwhelming agreement that
this is the right way to spend government money, and such agreement
will be hard to come by in the current competition for public money.
The older utilities were built by private enterprise, because they
thought people would be willing to pay for the service. This
requires giving or selling franchises to companies formed to provide
the service. It is not clear that private enterprise is as
adventurous as it was in the nineteenth century. The development
costs are likely to be hundreds of millions, perhaps even a few
billion.
Some social consequences.
Realization of the full advantages of automated delivery will cause
large changes in the operation of stores. For example, there might
come to be stores with no premises where goods are kept. The goods
bought at wholesale are delivered to public warehouses and delivered
to customers automatically from the warehouses. The store has
financial responsibility, because it buys the goods, prices them,
advertises them, and sells them, all on the basis of its ideas of what
the public will buy. Such a separation of the marketing function from
physical goods handling functions will increase flexibility and
competition and will ultimately give consumers greater choice.
1995 note: Mail order houses with overnight Fed Ex delivery have
met a certain amount of this need.
Send comments to jmc@cs.stanford.edu.
The number of hits on this page since 1995 December 22.