The goal of this proposal is to provide a less costly alternative to long haul trucks.
Beginning with the shipper, the shipper loads his freight onto a semi-trailer. The semi driver takes the trailer to a train siding where the trailer is loaded onto a self powered railroad car.
As an approaching train slows to about 50 miles per hour, the self powered car accelerates on the siding, then onto the main line, catching up with the train and hooking up with it. Then the car is pulled along by the train.
The train's steel wheels have significantly less rolling friction than the tires on trucks. A railroad car, by it's closeness to the car ahead, drafts it, significantly reducing aerodynamic drag. To move freight on a railroad requires about 1/3 as much energy as moving freight by truck. There is a significant potential for fuel savings and reduced pollution.
For a car to leave the train, the train slows to about 50 miles per hour, the car unhooks and is switched onto a siding. Any following self powered cars would need to use their power to catch up with the train and hook up again. It'd be helpful if a track switching mechanism could be devised such that a device on the car itself was able to select which of two tracks it was to go on.
Note that once on the siding, the car could catch a different train going in a different direction. It'd be somewhat like going from one freeway to another.
All of the preceding would be under computer control, including delivering the car to the destination siding. The self powered railroad cars would be unmanned.
Once at the destination siding, a semi would take the freight to it's destination.
Unlike trucking where there is a requirement for drivers to have down time, once on the tracks, the freight could be on the move much of the time. The railroads could have schedules such that transit times could be calculated with good accuracy.