Ideas for Freight Railroad Modernization

by David S.Lawyer More railroad articles by David Lawyer









8. SWITCHING (or Pickup and Delivery of Cars)






Copyright 2003 by David S. Lawyer. Feel free to make copies but commercial use of it is prohibited. For example, you can't (except to an insignificant degree) combine it with advertising on the Internet. Please let me know of any errors or suggestions for improvement.


This article was written by me 17 years ago, so some parts of it are not valid today (2003). But most of it is still valid. I'll revise it as time permits. I never tried to publish it and only showed it to a couple of people. But now it's on the Internet for all to look at.

Today, the U.S. Federal Railway Administration is proposing some of the ideas presented here, but the proposal doesn't go nearly far enough. See 5-Yr Plan Chapter 2: Intelligent Railroad Systems

The proposed ECP (Electronically Controlled Pneumatic Braking) is similar to what I proposed. PTC (Positive Train Control) is not clearly explained and it's not clear if it can be used to drive the train without the train crew doing anything but observing. It doesn't seem to discuss the labor union issues involved in this.

A good candidate for train control software would be utilization of the Linux Operating System. This possibility isn't discussed either.


Railroad freight transportation offers a potentially low-cost and energy-efficient means of transporting much of the nation's intercity freight. Although railroads have lost much of their freight traffic to trucks, changes in technology and operating policy could recapture much of this lost traffic. Since freight transportation in the U.S. represents about 8% of the gross national product, cutting the cost of freight transportation by say a third could save the nation about a hundred billion dollars each year. One way to significantly reduce freight costs is to drastically change railroad technology and policy so as to both reduce costs and improve service. While the savings in rail costs (and the shipper benefits due to improved service) would be substantial, most of the savings would occur due to the resulting diversion of freight from trucks to rail. This is because truck unit costs (per ton-mile) are currently several times that of rail --and trucking accounts for about 3/4 of the nation's freight bill (rail is about 1/10 with the remainder shared by water, air and pipeline transportation). [see Transportation in America (annual), Wash. D.C.]

At present, rail service is inferior to trucking because it is relatively slow (The average speed of loaded freight car is only about 5 mi/hr. [guesstimate by D. Lawyer]) and unreliable (1/4 of shipments are 2 or more days late [see "The Railroad Situation", report No. FRA-OPPD-79-7, U.S. Gov't Printing Office, Mar. l979. p.48]). In addition, railroads have lost the ability to efficiently handle shipments less than a carload in size [see "The Railroad Situation", op. cit., p.337]. Intercity trucks, (the large "semis") like trains, often operate both day and night with relief drivers. Thus the average speed of such trucks is perhaps 25 mi/hr taking into account the time truck drivers spend stopped en-route as well as slow-downs due to grades, traffic, signals, speed limits etc.

Since the specific rolling resistance of a train is only about 10% of a highway truck, it turns out that the optimal speed to use for trains is usually faster than truck. It is indeed ironic that the actual average speed of rail freight should turn out to be so much slower than truck. Of course, rail freight cars must suffer delay when they accumulate to be made up into trains, but one would expect the application of modern technology to result in an average speed of a few times higher than the present 5 mi/hr speed for the typical 600 mile trip.

The "railroad problem" is both social and technical so that technology alone cannot solve the problem. Railroad labor unions (and the reluctance of management to vigorously pursue change) have resulted in wasteful and costly operating policies such as paying a train crew of 4 or 5 men a full day's pay for about 100 miles of "work" when 1 or 2 men could handle the job and take the train a few times as far in an 8 hour shift. Even track laborers are paid a basic wage of about $35,000 per year plus 40% for fringes. [see Dick, Merwin "M/W Review and Outlook ...", Railway Track and Structures, Jan. '85, p.18. See also "The Railroad Situation", op. cit. pp. 211-2]

While the rationalization of labor policy would significantly reduce costs, a revolutionary improvement in service also requires freight cars which have computer control of both couplers and brakes. This means a complete redesign of the 100-year-old archaic braking and coupling system. The new couplers would also automatically couple electrical circuits to provide electric power and control/monitoring signals to (and from) all cars in a train. The self-braking of coasting cars would be controllable from both fixed and hand-held computers which makes possible new and rapid methods of switching and sorting cars. Although such basic changes would be very costly, the benefits might significantly exceed the costs. Before explaining the details of the new technology (and policy) here is a list of some of the improvements it would engender:

1. Greatly speed up and reduce the cost of sorting cars in yards. At present, cars are delayed nearly 24 hours at a typical yard [see "Railroad Classification Yard Technology", report No. FRA/ORD- 76/304, Jan. l977, pp. 42-3, tables 9,10].

2. Give the locomotive instantaneous and full control of braking. The existing brakes do not all apply simultaneously (the time lag may be as much as a minute) and as a result, slack action may damage freight. To avoid slack action, trains often operate under power with their brakes on. This wastes energy. On steep downgrades the locomotive engineer cannot reduce the braking effort of the cars of his train since the existing brakes are not fully controllable.

3. Speed up and reduce the cost of pickup/delivery of cars to customer sidings (either private or public sidings) including spotting of cars in a position for loading/unloading.

4. Increase the reliability of cars by monitoring their condition.

5. Permit removing and/or adding cars to a moving train without stopping the train. However, cars to be added would need to first be imparted velocity by a locomotive, downgrade, or catapult..

6. Make it feasible to operate most linehaul trains with most cars in the desired order.

7. Make it economical (due to reduction in switching/sorting costs and delays) to use smaller cars with less axle load. This would both reduce the damage to the track caused by the existing high axle loads as well as permit the use of two-axle cars which could recapture small shipment traffic from trucks.

8. Make feasible the use of helper locomotive units to speed trains over grades (since the helpers could be added/removed without stopping a train).

9. Permit trip logs to be recorded on magnetic tape in each car so as to assign responsibility for damage and/or delays. Only the railroad customer would have access to the tape which would record jolts and the trip itinerary.

10. Permit distributed locomotive units in the train thereby reducing longitudinal forces in the train. The lead locomotive would control the other slave locomotive units uniformly distributed within the train.

11. Eliminate the damage due to the excessive coupling speed of cars in yards.

12. Reduce derailments by detecting rolling/rocking of cars as well as severe bumps.

13. Control access of people and thieves to cars by use of burglar alarms in the cars. The use of empty box cars and refrigerator cars (warmer due to insulation) to transport people would become more feasible.

14. Provide electric power to cars for operation of brakes, couplers, hopper doors, and possibly power steering. The existing self-steering trucks result in both a high rate of wheel wear and wasted energy due to hunting.

15. Allow voice telephone communication throughout the train. This would be used by passengers and repairmen using portable plugin phone-sets.

16. Permit non-stop meets of trains. Trains going in opposite directions (on a single track line) would pass each other non-stop on short double track sections (passing sidings)..

Unlike the electricity-less freight cars of today with their 100-year-old braking and coupling systems, the new cars would be amply supplied with electric circuits running the entire length of the train. These wires and cables (some coaxial) would convey both electric power and digital/analog messages for a variety of purposes. Each car would have its own storage battery which would be charged by electric power supplied from the locomotive units. The couplers would be truly automatic and would uncouple upon command of the locomotive or wayside computer. The couplers would automatically connect (and disconnect) electrical plugs/jacks thereby providing electrical circuits running the length of a train.

Each car would have a small computer which would monitor and control the operation of the car as well as communicate with other computers. A car in a train would communicate with the locomotive computer while a car not in a train would communicate with other computers (either hand-held or located in buildings along the wayside). The communication channels between rolling stock and the wayside would include the rails, a trackside cable, or radio.

Each head locomotive computer would communicate with and control the computers in the cars of its train. In some territories, this locomotive would itself be controlled by computers located along the wayside. In other territories, where low traffic density does not justify the computer control of train motion, the train would be under the control of the train driver (or operator).

A major roadblock to the rapid sorting of cars is the fact that the railroad car is not self-propelled. Locomotives or gravity must be used to move it during sorting. If each car could be provided with its own motor, sorting would be greatly facilitated but motors are both costly to buy and to maintain. In addition, the utilization of most such motors would be low since cars spend considerable time being loaded/unloaded or in storage due to lack of demand. If such motors were to be used, the lowest cost solution would be electric traction motors due to their low maintenance cost.

Unfortunately, such electrically motorized cars could not be used as independently powered units unless railroads are electrified, preferably using a standardized system of voltage/frequency. Rail electrification is not well developed in North America and in Europe it is not a standardized system (Some systems are AC, others DC, and voltages vary.). Even with standardized electrification, it is not economical to electrify lines with low traffic density. Thus gravity and locomotive powered sorting methods would still be needed even if railroad electrification becomes a reality in North America. Therefore, self-propelled cars will not be used in the proposed system but provision should be made for their use should rail electrification become extensive in North America.

A railroad freight system with modernized cars has been possible for many years but in the past, computers were both more expensive and less reliable. Today, with much cheaper computers, the computer control of train and car trajectories is becoming economically feasible (if not mandatory if railroads are to provide a high level of transportation service to the nation). To ensure reliability, redundancy will be required with computers comparing results before commands are executed and with feedback to make sure that commands are being executed properly. With all computers programmable, changes in control and operating policies could be readily made.

Major changes, such as this proposal, have historically been imposed on the railroad industry by the government. For example, the introduction of the automatic air brake and the automatic coupler was imposed (despite the protest of many railroads) by the federal government almost a hundred years ago. [see Stover, John: Americans Railroads, Univ. of Chicago Press, l961.]. Modernized cars (along with radical changes in labor and operating policy) will probably need to be imposed by the federal government. One alternative is government ownership of railroads. More about these complex questions will be discussed later in the section on "Institutional Issues".


The present train braking system results in slack action which damages freight. In addition, it is not really controllable since once the brakes are applied on a downgrade it is not possible to reduce the braking effort without the lengthy (and often dangerous) process of fully releasing the brakes first. To enable cars to coast in yards requires manual time-consuming bleeding of the brakes on each car. It may take up to an hour for a locomotive to pump up the brakes of a train (with its compressor) so that the train may get underway, although use of compressed air at stations often speeds this up. The proposed new car would not only eliminate these problems but would permit wayside control of braking.

With all cars communicating with the head locomotive via a train circuit, an electrical (digital) command sent from the locomotive would apply all the cars' brakes simultaneously, thereby eliminating the problem of slack action. In addition, the computer of each car would prevent wheel locking by ensuring that the braking force of a car is proportional to its current weight. A small magnet attached a each axle would send pulses to the car computer which would detect and correct wheel skidding as well as serve as an accurate odometer. This would significantly reduce the coupler forces during emergency braking and thus reduce the chance of derailment.

While all cars would have their brakes electrically controlled, different types of brakes could be used in different cars. The existing cars could retain their air brakes while adding a small air-compressor (electrically operated) to each car to furnish the compressed air for the brakes. Such an electro-pneumatic brake system is also a possibility for use in newly constructed cars. One advantage of air brakes is that the energy stored in compressed air may be converted into useful work almost instantly while the energy from a storage battery cannot all be used in single brake application.

Another possible braking system for a car is electric-hydraulic. A small high-speed electric motor would be geared to as to produce a high force on the piston of a master hydraulic cylinder. The master cylinder would supply hydraulic pressure to operate the wheel cylinders as in an automobile. Feedback by a pressure transducer would permit the car computer to both monitor and control the fluid pressure and hence the brake shoe force. To speed up the brake application, the gearing between the brake electric motor and the master cylinder could have two gear ratios. When the brakes are first applied, a low force (at high speed) would be applied to the master cylinder piston so as to bring all brake shoes into contact with the wheels (or rotors if disk brakes are used). Then the gears would downshift so as to produce a higher fluid pressure for a strong braking effort.

Since each car has a battery to drive the braking electric motors (or air compressor), the brakes of a car could be controlled even when a car is coasting by itself (and is not in a train). This feature would permit control of car braking from the wayside and make possible radical changes in yard technology so as to process cars through yards in a small fraction of the present time. In locations (such as near customer sidings) where there is no wayside computer system to control the braking of individual cars (out of a train) a car could have its braking plan preprogrammed by a previous order from its last locomotive. This would permit a train to release cars from it (by a command from the locomotive to uncouple the appropriate couplers) with these cars braking to a halt under a preset program without control from the wayside. Another way of controlling the brakes in such circumstances would be by a person standing near the wayside holding a hand-held computer terminal.

Should a train coupler fail and a train spilt in half, the existing air brake system will apply a full braking effort to both halves of the train. The computerized system would improve on this by braking the lost part of the train at a higher rate so as to maintain separation. The engineer (driver) would be notified as to where the break occurred.


The car couplers would, like the brakes, be controllable either from the locomotive or the wayside. The force to "pull the pin" (uncouple the coupler) would be supplied from a "coupler (electric or hydraulic) motor". Using transducers to measure coupler forces would insure that no attempt is made to uncouple while the coupler is in tension. The coupler would consist of two subsystems: the mechanical coupler and the electrical coupler. The mechanical coupler would transmit the traction and braking forces between cars of the train while the electrical coupler would connect the train circuits. The design of the electrical coupler is a challenging problem since the connections must be reliable and operate in all kinds of weather.

One possibility is to enclose the electrical plugs and jacks (connectors) in watertight metal boxes which would protect these connectors from the weather. When two cars couple, these metal boxes would also mechanically couple together but the electrical connecters would not yet be coupled. Then a door on each box would open inward, thereby creating a weather-tight passageway between the boxes for the electrical connectors to couple in. Driven by electric motors or hydraulic cylinders, several electrical connectors would then couple. The forces applied to them would be low enough so as not to break the connectors if they refuse to couple due to a bent pin, misalignment, etc. If this force was above normal, a request for servicing would be issued. Of course, a more urgent request for servicing would be issued if an electrical connecter failed to couple at all. About half of the connectors would be reserve ones which would not normally be used (although they would be subject to continuous testing). Furthermore, spare train circuits would be provided. Thus, the failure of an electrical connector or a broken wire would seldom disable a train since switching would automatically disconnect defective portions of circuitry and bypass them with operable circuitry.


The most important single benefit of the proposed system will probably be a large reduction in delay in yards. In order to describe the proposed switching methods to be used (both in yards and on sidings) some new terminology is needed:

1. String = a string of rolling stock coupled together. The existing term is "cut" meaning cars which have been cut off from a train. The term "string" is more descriptive since a string may be created in ways other than just cutting some cars out of a train. Examples of a string: a single car, several cars coupled together, or an entire train. A string is usually an isolated string and becomes only a substring (or possibly a block) when it is coupled to another string. A string is thus tantamount to a vehicle.

2. Block = a "purposeful" substring. A block need not be isolated from other rolling stock and is often just a contiguous group of rolling stock within a train which has the same next destination (but not necessarily the same final destination).

3. Train = a string which contains one or more usable locomotives.

4. Gap = the unoccupied space (along a track) between two adjacent strings ranging from a few centimeters to a kilometer or more in length. If the two strings are moving the gap is the headway between the rear of the leading string and the front of the trailing string.

5. Stream = a stream of strings = one or more strings flowing (i.e. rolling) down the same track in the same direction, one behind the other but separated by gaps.

6. Cut = to uncouple a coupler within a string, thereby breaking the original string into two new strings. Also, the location where a string is cut.

7. Bunching = the elimination of gaps by joining (coupling) strings together. Similar to the concatenation of strings in computer programming.

8. Separation = the opposite of bunching = the creation of gap(s) by cutting a string at one or more locations. This partitions a string into substrings.

9. Diverging = the flowing of a stream into a fork (or series of forks) where some strings take the left fork and other strings take the right fork in the road (here a railroad). The point of the fork is a railroad switch. This is similar to autos in the right-hand lane of a freeway either selecting or rejecting to exit the freeway at an off-ramp. When there are a number of forks (switches) in series, a string may be diverged onto one of many possible output tracks.

10. Merging (or Converging) = the opposite of diverging. Imagine a diverging operation is about completed and let time run backwards so that strings (or vehicles) flow in the reverse direction. This is a merger. A simple example is a stream of autos in the right-hand lane of a freeway merging with the stream of autos entering the freeway via an on-ramp. However, in the more general case, several streams may merge together to form a single output stream.

Cars are sorted by merging and/or diverging. One may define verging as either converging (merging) or diverging. Before verging can begin, streams must be generated. A stream is created out of a moving string by differential acceleration. I.e. a string is cut (a coupler within the string uncoupled) and the resulting two substrings are accelerated at different (differential) rates to achieve separation.

One method of separation is to push the input string over the crest of a hill (i.e. a hump) and cut the string into substrings as it passes over the crest with differential acceleration (due to gravity and the changing gradient) providing the separation as the strings roll down from the crest. Another method is to create a stream by kicking. To do this a locomotive pushes the original string which is cut in two just before the locomotive deaccelerates. The leading string created by the cut then coasts out ahead resulting in separation. Both the above methods are currently in use but cutting requires a man to pull the coupler lever. A third method is to cut off a rear section of a moving string and deaccelerate this rear substring (possibly by applying its brakes). This new method will be called "rear release". The methods of kicking and rear release can be repeated to create a stream of several substrings. Kicking of a long input string may use up much energy and time in the repeated sequences of acceleration and deacceleration of the string in front of the locomotive.

Most of the sorting of cars today is done by diverging. Controllable cars will permit a shift of emphasis from diverging to merging. The results of a merger can be a train which is not only ready to depart, but is also moving (hopefully towards its destination). The new controllable cars would also greatly speed up the verging of cars.


Today, there are two basic types of yards: hump yards and flat yards. The hump yards have a hump as described above where a string of cars is pushed over the hump by a locomotive. This achieves separation and the stream enters the root of a tree type network where diverging takes place. Each string created by cutting at the hump winds up on one of many classification tracks below the hump. Each classification track usually holds only one block of a train. To create an outbound train, a few (or several) classification tracks are then pulled (a locomotive removes the blocks from them) and the blocks are put together end to end to make up an outbound train. This is tantamount to the merging of these blocks.

A railroad yard is to railroad cars as an airline terminal is to air passengers. Air travelers change planes and cars change trains but the cars get delayed several times as long (almost 24 hours is typical). Yard delay is a major cause of rail's poor level of service. Cars with remotely controllable brakes and couplers make possible the rapid sorting of cars. Combine this with better scheduling of trains and delays in yard would be reduced to a small fraction of the current delay.

Yards should be places where cars rapidly switch from one train to another rather than places where cars accumulate and wait to be made up into trains. Trains should be scheduled so that several trains (including local trains) arrive a yard at about the same time. Then these trains would rapidly interchange their cars with each other and quickly depart the yard although a small percentage of the cars would need to remain in the yard for repairs, to catch a delayed train departure, to be delivered to a local customer, etc.

One factor in yard delay is the inspecting of both arriving and departing trains. Since the car and locomotive computers would have a record of car defects obtained from monitoring the car's condition, the requirements for inspecting cars in yards would be reduced to a minimum. Whether or not a defect was repaired in a yard would depend on both the seriousness of the defect and whether or not the car was ahead of schedule. If feasible, repairs and maintenance should be postponed until a car is both empty and not in short supply. This policy would improve the travel time reliability of loaded cars.

Many trains that now run through intermediate yards without discharging or picking up cars would start to service these intermediate yards. This would provide better service by decreasing "usable" headways. Since trains will get through yards rapidly (usually in a fraction of an hour) this policy will not unduly decrease the scheduled speed of such trains. For example, if a loaded car in a yard was headed west it would normally just take the next non-local train headed west. With some exceptions, line-haul trains would keep their cars well ordered (blocked) within their trains.

As a train arrives at a modernized yard, its cars would first diverge. Then by merging of new cars into the continuing portion of the input train, a new output train is created. The delay to cars of the input train in getting through the yard would be measured in minutes instead of hours as is the case today . Two examples will be now given:

A "Crossroad" Example: Consider a yard located at the intersection of a main line and a branch line. As a mainline train approaches the yard it separates into subtrains with each subtrain powered by a locomotive (which was formerly one of the locomotives distributed within the input train). This creates a stream where each string of the stream is itself a train and thus has sufficient power to maintain velocity. As this stream of subtrains gets closer to the yard, each subtrain (if required) separates by the method of rear release into strings. The resulting stream enters the yard and diverges for the purpose of releasing the cars destined for the yard as the terminating destination (including the cars destined for the branch lines).

As the original train was approaching the yard and separating, all available locomotives in the yard were being used to accelerate cars in the yard which are to be merged with the arriving train. The trajectories of these cars (in some cases strings) will fit nicely into the gaps of the continuing input stream (just after diverging). Finally, the output stream is bunched (all strings couple together) while in motion and the resulting output train continues on its journey.

Thus the input train never even had to stop in order to release and add cars to it. Furthermore, the cars may be added or released anywhere in the train (provided there is enough locomotive power to achieve the proper trajectories). In many cases, the mainline train will slow down (usually by coasting in order to conserve energy) as it approaches the yard so as to permit the limited power of the yard locomotive to successfully accelerate the cars to be added to the train. In other cases, a "yard" locomotive may remain attached to the cars which it accelerated and become part of the mainline train output from the yard.

Diverging will often occur at a single switch near the entrance to the yard. However, merging may use more than one switch in order to avoid interference during the acceleration of cars from the yard to be merged.

Another example is where a yard is located near the intersection of two main lines and handles trains from these lines. An ideal scenario is for four trains (from all directions) to arrive the yard at about the same time and to interchange cars with each other. In the ideal case, each incoming train would contain four blocks: 1. a continuing (in the same direction) block. 2. a left turning block. 3. a right turning block. 4. A block destined for the yard with cars destined for customers serviced by the yard (often within several miles of the yard).

As each train approaches the yard its blocks are separated and diverge, each block going to a different track. Then all the blocks for a certain output direction are merged together to form an outbound train. This train contains, in addition to the three blocks obtained from inbound trains, a fourth block of cars originating at the yard. Next another outbound train for another direction is formed by merger. If the input blocks are themselves well blocked (the cars within each block are well ordered) then the output trains will have their cars well ordered.

There are obviously many other verging scenarios for yards and much more remains to be done to develop verging strategies. There is also the problem of track topology, which must take into consideration the cost of additional land, if needed. How much use should be made of humps for verging operations? Each yard presents a different problem but a first step would be the preliminary design of a few typical yards.

8. SWITCHING (or Pickup and Delivery of Cars)

This is the movement of cars between customers and the nearest (usually) yard. It represents either the start or the end of a car trip. Customers are located either at private sidings or use public sidings sometimes known as "team tracks" since a team of animals was once required for hauling freight to/from a car on such a siding. Four or five man crews usually operate these switching trains. While a smaller crew would often be feasible, labor agreements seldom permit it on major railroads. Unfortunately, one man operation with the present technology is quite a problem since a second man (in addition to the engineer) is needed to uncouple cars because the brake lever must be held up as the locomotive pulls the cars away from the cut. When the locomotive is pushing cars and the track ahead is not visible to the engineer, an additional man is needed to look down the tracks and render guide service for the engineer. In some circumstances it is convenient to have three or more men such as when switches need to be manually thrown in addition to the above mentioned tasks.

The new technology will often permit one man operation. With a remote controller (computer terminal) in his hand the engineer will be able to stand on the ground at a suitable vantage point and command the performance of switching operations with the computer taking care of the details such as throttle settings, braking rates, etc. At times, when the locomotive is pushing cars, he may elect to ride on the front car and remotely control the train by plugging his remote controller device into a jack located on the end of the car. Safety straps would permit him to use both hands to operate the remote controller. If he doesn't ride the front car, he could place a warning beeper (or flashing light at night) on the lead car as a safety measure.

Much of the process of switching can be done under direct computer control (with some intervention by the engineer). The exact position of all the cars involved must be known including the customer cars which are to be picked up. The customer could assist by notifying the railroad of the position of his cars, based on painted markings on his loading dock. Similarly, the customer would specify exactly where he wants an arriving car placed.

The engineer would issue commands to the computer to start the switching operations and then observe their progress, intervening if required. Situations requiring intervention include: 1. Incorrect locational information of cars. 2. Track obstructed by vehicles, animals, etc. 3. Hardware (computer, rolling stock, switches) malfunction. 4. Software errors.

The engineer would manually throw track switches that have not been converted to electric operation. A low-cost, possibly flashlight-battery operated switch-thrower, needs to be developed for such seldom used switches.

A wide variety of track network topology exists for switching tracks and a survey of typical track configurations is needed before specifying the switching policies to be used. If a customer siding is on a through track, and if the customer has all cars on this track ready to depart, then the switching train may simply couple onto the waiting cars and push them out of the siding. As the switch train departs it may drop off cars from its rear that are destined for this siding, with the brakes preprogrammed to stop at the correct spots. The engineer would usually ride the front car of this train with the locomotive located somewhere in the middle. Upon picking up new cars he would have to walk (or run) past these new additions to get to the new head of the train.

In some cases it would be advantageous to use two small locomotives on a switching train. Each locomotive would have a one man crew. While one locomotive stayed attached to its switching train, the other would separate from it and rearrange cars on sidings so that they could be more rapidly picked up by the switching train.

The new technology would make it safe to perform maneuvers which are now against the rules. One of these is called a "flying switch" or "Dutch drop". This is the case where a locomotive which has cut itself off from the string of cars it was pulling, speeds up and gets to a switch well ahead of the string. After stopping beyond the switch (after passing through it) the switch is thrown and the locomotive backs through it. After throwing the switch again, the coasting string catches up to the switch and rolls past the waiting locomotive. Then, throwing the switch again, the locomotive chases the string and couples onto it. the final result is that the locomotive has reversed its position from one end of the string to the opposite end.


With a computer in each car, the performance of the car may be readily monitored. In order to conserve energy from the battery, the car computer might be switched to a low-power standby mode when nothing special is taking place. The full computer would be activated by a signal from the locomotive, from an alarm condition detected in the car, or by a digital clock in the car's computer which would always run during a car trip.

Each car would be equipped with various sensors to monitor the physical state of the car. An important sensor is the bump detector which would both record high intensity bumps and jerks which often damage freight. Lower intensity bumps and vibration would be periodically sampled by the locomotive as a check on the condition of the car. This would detect wheel flat spots and defective suspension systems.

Sensors would also measure the condition of the brakes (such as the amount of wear as determined by the state of the automatic slack adjuster) and wheel bearings (temperature). An estimate of wheel wear may be obtained by comparing the reading of axle odometers with the known distance between fixed markers located along the wayside.

The goal would be to detect impending problems so that they can be corrected without the movement of the car falling behind schedule. Most repairs would hopefully be made when the car is empty so as not to delay shipments of freight.

Each car would have a burglar alarm system to reduce theft. Alarms from this (such as attempts to open the car doors (or hatches etc.) would also be recorded in the trip log. Alarms would also be transmitted to the locomotive, the wayside where possible, and in some cases to the railroad customer when the car is on his siding.


The car computer would compile a trip log which would measure the quality of service to the customer and help pinpoint who is responsible for delays and damage to freight. The locomotive would also keep a log which would record much of the same information as recorded in the car. This duplication is not only to improve reliability but also to discourage falsification of the log by the railroads and their employees.

The railroad customers would directly obtain the car log when the car arrives at its destination. The log would be accessed by a long code or password known only by the sender and receiver of the freight car. In some cases the log would be read out of protected memory of the car computer but in other cases it would be recorded on magnetic tape (perhaps a conventional type cassette) which would be physically removed when the car reached its destination. To protect the tape against tampering, the car tape recorder would not be capable of reversing, rewinding or erasing. The tape compartment would be securely locked with access only available to the railroad customer. The opening of this compartment would also be logged in protected memory. Freight transportation has a long history of all too frequent cheating of various sorts and thus such precautions (and more) are necessary.

The log would record the time which the car arrived and departed yards as well as other major point en-route. It would also record the time and location of severe bumps and jerks and in some cases record temperature and humidity. A car odometer would furnish the log with mileage data which would be supplemented by geographic locational reports from the locomotive.

The contents of logs would vary with the type of car and the requirements of the customer. For example, shippers of low-value bulk commodities such as sand and coal may not be much concerned about bumps and jolts (unless they have been so severe as to cause the cargo to spill). Some customers may prefer a printed log which would be mailed to them with their bill, rather than a log on magnetic tape.


The train computer system could help prevent derailments and also reduce the amount of damage caused by a derailment. Some derailments are caused by car rocking which can be detected by sensors on the cars. Trains would detect and log serious defects in the track which need to be repaired. A bad bump detected by the head locomotive would be reported on by each car as it passes over it. A bump which gets worse as more cars pass over it may indicate progressive track failure and in some cases a braking command would be issued to slow or stop the train.

It may be feasible to detect the actual derailment of a car. In this case, the policy could be to cut losses of other cars by uncoupling the derailing section of the train, braking the section behind it while the front part of the train would pull clear of the derailed car(s) and then stop. More research is needed on the optimal behavior of a train during a derailment where uncoupling of cars and braking different parts of the train at different rates is possible.


Today, nearly half of the freight cars travel empty. Empty box cars offer a potentially low cost passenger transportation system with a low level of service. Since the Civil War there has been a significant amount of riding of freight trains by non-paying passengers. These passengers are mostly poor and include farm labors, illegal aliens, homeless persons (disparagingly call hobos or tramps), and even students. Most of this riding is illegal although many railroad are lax in enforcement of rules and laws against it. With computer monitoring of car doors and telephone communication with the cars, it will be easier to enforce rules regarding passengers but the rules could be changed to permit riding.

New boxcar doors would be capable of being closed and opened from the inside. One way to provide air and light for passengers is to enable the doors to be locked in a partially open position. Another possibility is to equip boxcars with skylights and adjustable air vents. Pastel interior colors would reduce the need for skylights. Some cars could have electrical outlets for passenger use which would either supply say 12v DC from the car battery or 120V AC from the locomotive. The main use of these would be for the lighting the interiors at night using lights supplied by the passengers.

Telephone jacks could be provided which could be used with the low-cost headsets used on miniature radios/tape-players. The passenger would need to have (or borrow) such a headset in order to communicate with the train driver (or wayside station) but the microphone could be built into the car.

It would probably be uneconomical to heat ordinary boxcars as they have no insulation. Refrigerator cars are insulated and thus may be feasible to heat. Even so, most trains would have no empty refrigerator cars and thus most passengers would have to rely on warm clothing, sleeping bags, and mattresses (air or foam) for warmth.

To ensure a pleasant journey, enforceable rules must exist for passenger behavior. For example, in some cars smoking would be prohibited. During nighttime hours, passengers would not be allowed to make any noise which interfered with anyone else's sleep. Even during daytime hours, loud noises from audio equipment should not be allowed unless there was unanimous consent. Passengers who smell badly would not be permitted to board a train.

The enforcement of the rules would be in part by telephone connected either to the locomotive engineer or the police in a nearby city. In case of violence or serious violations, the car could be removed from the train (without stopping the train) into the hands of waiting police. It may be desirable to have built-in antennas in each car which could pick up signals from citizen-band radios carried by some passengers. This would provide an alternate channel for emergency communication.

Since boxcars are not equipped with toilets, a major problem is how to dispose of human wastes. For today's illegal boxcar riders it is not too much of a problem since most of them are men and can easily urinate out of the car doors. Since today's trains frequently stop at sidings for meets, there is often the opportunity to leave the train for a few minutes. While the new technology would make freight train riding more attractive by greatly reducing the number of stops, it would at the same time aggravate the waste disposal problem. With freight train riding legal more women may be expected to ride boxcars posing additional problems.

Assuming that little harm is done by disposing of urine on the roadbed, boxcars could be built with a small hole in the floor for this purpose. A funnel (perhaps recessed) would fit into this hole and passengers would be expected to sprinkle water (from their canteens) on this funnel to clean it after using it. On long trips the defecation problem has three possible solutions: 1. The railroad would provide small (perhaps only a foot high) portable toilets for each car. 2. Passengers would carry plastic containers with them. 3. If required, a passenger would make a short stopover for the purpose of using a wayside restroom. While waiting for the next train, he might be able to usefully utilize his time eating in a nearby restaurant, purchasing food and supplies, refilling his canteen(s), etc.

The passengers, in addition to paying a small fare for transportation, would be required to assist the railroad in case of emergency (accident, train stuck in snow, attempted theft of train cargo, etc.). Before departing on a trip, a prospective passenger would first make reservations and buy his ticket. Use of public (or private) video terminals for this purpose would automate the reservation process and help reduce costs. A passenger would then board his boxcar and report his presence to the locomotive engineer over the telephone line. He would also report the number of passengers in the car. There would be no need to collect tickets since the locomotive engineer would have the passenger roll (manifest) of those who had purchased tickets and would check them off (using a computer terminal) as they gave him their names over the train phone line. The engineer would also attempt to talk on the phone to anyone not on his manifest to determine the cause. Persons boarding without a reservation would have to pay both the fare plus a reasonable fine.

Today, even with all the dangers, illegalities, and lack of amenities, there is still a significant amount of illegal riding of freight trains (The typical train may have several illegal passengers aboard.). With the improvements in the level of service and safety proposed here, a significant increase in ridership is to be expected. This would help reduce the nation's passenger transportation costs and provide safer and better low-cost transportation for poor people.