Dual Mode

by Professor Dwight Baumann
Carnegie-Mellon University, Pittsburgh, Pennsylvania

The article was originally published as Chapter 3 of Research Report 5 entitled Advanced Urban Transportation Systems, by the  Transportation Research Insititue, Carnegie-Mellon University, Pittsburgh, Pennsylvania, in 1970.

The advanced urban transportation systems now quite universally called "dual-mode" have many antecedents. The basic concept of what now has the generic name "dual-mode" centers around the hybridization of an automated, individual, transit system and the private automobile, while retaining most of the features of both parents. The two modes relate to the system's capability of either being driven by a human driver in an automobile mode, or by a control system in an automated transit mode. A dual-mode system uses the same right-of-way for both transit and auto applications. Since the acquisition and use of right-of-way is the most expensive and potentially contentious aspect of a transportation system, this dual use offers a compromise solution between the two political adversaries, the transit and the highway interests.

The term dual-mode has been used predominantly to describe an urban transportation system as contrasted to either an interurban or a local circulation system for high density, core-city, or airport use. The primary application anticipated for the commuting trip is in highly congested areas. Usually the term has been restricted to electrically powered systems that are thus not direct polluters of the atmosphere. In the automated mode, the vehicles can also be parked automatically, thus developing considerable flexibility of storage, ownership, and availability. Taxi, bus, auto rental, and privately-owned versions of a dual-mode vehicle may differ only in amenities such as cleanliness, choice of co-passengers, access to golf clubs stored in the trunk compartment, etc., and will be billed at the costs indicated by the amenities provided. The basic concept of automating ground transportation systems is not new. Actually there is perhaps very little that is really new in the entire transportation field if one considers only the physical concepts of the hardware systems.

This should not be too surprising as it has been estimated that perhaps one fourth of all the patents issued in the U.S. have possible relevance to transportation. Something on the order of 20% of the GNP is related to transportation. And the majority of the trips and the trip miles are urban, so we should expect that the major unsolved problems of an advanced urban transportation system, the topic of this conference, should be on a "systems level" of complication. Necessary breakthroughs in the urban transportation may possibly be related to innovation in the financial, political, jurisdictional, or management portions of the transportation system, while using technology that is comparatively close to the present state-of-the-art for automation, propulsion, structural, and vehicle technology. If this is true, then we have a significant answer for that large portion of our society that would like to believe that the problems await an important invention of a new gadget.

The laser, linear motors, air cushions, fuel cells, etc. are all candidates in the minds of many for that scapegoat, the required breakthrough. The monorail has been the required breakthrough machine for many generations only to always be found wanting in some aspects. Perhaps that mythical legal character, the reasonable and prudent man, Mr. Average, does dream of sleek vehicles that fly through the sky. Isn't in fact, everyone a transportation expert, and doesn't he dream of banking around curves and stopping at just the right place—where he wanted to go. Each time the monorail dream has been investigated, however, some of the same problems are revealed. First of all it doesn't go to the desired destination. It stops at intermediate stops and thus no matter how high the top speed, it averages below 25 mph for stations one mile apart and less than 20 mph for stations one-third mile apart (1). If a network is to be traversed, the waiting time at the stations increases, and it often entails considerable walking to bring the average travel speed down to something on the order of 5-10 mph. All this for that sleek, streamlined vehicle.

On the technical side, the freedom to bank on curves also allows buffeting from wind and passing vehicles, and must be separately controlled, only to find that more boggie wheels are needed along with dampers and other paraphernalia. Often it is decided that an underslung system is really more costly and that an over-riding system is indicated. The over-riding system, however, is similar to a railroad. And there we are, full circle with no breakthrough after all. Coming to grips with a problem the magnitude of the urban transportation problem, one must include as a basic assumption that any change must be evolutionary in nature. The small degree of revolution needed, however, to create any change has a threshold to overcome which is proportional to the size of the problem. Overcoming this threshold will certainly be easier to accomplish by making use of as much of the proven prior art as possible. Every systems designer, like every chess player, knows that it is important to plan several moves ahead in order to be sure that the contemplated procedure is a correct one. The evolutionary staging must be viable all along the way. And the most difficult of the decisions are associated with maintaining flexibility to accommodate changes in the perception of the problem, or in the actual environment of the solution.

The Automated Auto

Because the dual-mode systems are usually proposed for intracity transportation, they are not often related to their predecessor, the automated auto. The basic concept of automating the auto is not new. Similarly, autos that ride on a rail have long been used by the railroads. Railroad automation has been proposed and even tried, usually resulting in labor problems; i.e. by removing the motorman, a policeman would be required. Having all the ingredients is not the same as making a cake. Being able to automate the auto has long been considered an important goal. In the 1950's and early 1960's several electronics and automobile manufacturers developed prototype automobile guidance and control devices. The General Motors Autoline concept (2) and the RCA and Philco wire follower systems were conceived as intercity systems. The national preoccupation with the interstate highway system, the human proclivity for nostalgic concepts of travel and leisure, and the fear of the potential boredom that would be associated with the expressway all led to the consideration of the first automated auto system for long trips. Most of the automated systems were soon realized to be too expensive for long distance use because the expected traffic densities were initially quite low. Because startup could not be completed everywhere simultaneously, any system that required auto modification would not have enough converted autos at any one time to justify a dedicated lane.

Thus it was usually decided that an intermix of automated and nonautomated vehicles should be allowed. The presence of this mix of vehicles required that the drivers of the automated vehicles remain alert even though ordinarily they would not need to operate any controls. The presence of both driven and driverless vehicles required that substantial headways be maintained to account for the reaction time of drivers and monitors of the automated vehicle. The General Motors group determined that regularizing the flow would account for the major portion of the potential improvement created by an automated system. The magnitude of the improvement was determined to be only on the order of a factor of two increases in lane capacity. To obtain this factor of two, both the roadway and a very large number of vehicles would need to be equipped with electronics, and all this at an unknown and untested level of safety. It is indeed unfortunate that the automated auto experiments were not continued sufficiently to obtain data for actual application, as these results would be useful in deciding the redundancy requirements for electronics needed in a mass-produced guidance system.

One conclusion of the auto companies in earlier automated auto studies was that by increasing the amount of control equipment supplied to the user and the highway, its exposure to product liability suits would increase substantially. The official G.M. position on why it did not proceed on the automated auto experiments was exactly this concern, that it could not overcome the legal liability unless the government stepped in and defined liability limits that would reduce this exposure. This concern has had considerable discussion during the last ten years, and certainly the risk of exposure has increased due to recent court findings and increased public concern with highway safety.

The liability problem associated with automated transportation is certainly of prime importance in the development of any advanced urban transportation system. The Nader era has decreased rather than increased the amount of boldness that equipment producers are willing to apply to developing new components. Certainly the liability problem has several solutions, each or all of which must be engaged in the development of dual-mode. Legislative action limiting the liability of the system operator, the designer, and/or the manufacturer to a specified degree is one solution. This would be similar to the limited liability associated with international air transportation. A second approach would be to establish a massive insurance underwriting program that would evaluate the risks and add the cost to the user charge. Certainly if the new system is to be self-sustaining it must consider liability for property damage and personal injury as part of the operating cost; whether the designer, the manufacturer, the operating company, the insurer, or the government holds the funds needed for liability payments. Either limiting the liability or subsidizing it passes the cost on to the user with the possibility of inequity, depending on how many benefits can accrue from a system that is both public and private.

Since the automated auto experiments of the 1960's, there have been proponents of the idea that the automobile should evolve toward the complete automation of the driver's function. They argue that the driver's functions can be taken over by machines, one function at a time, until at last the driver is no longer needed. They suggest that first speed control, then steering, then control, etc., could be implemented. The improvement to an expressway that is already grade separated is assumed, when the primary problems are the public's concern where an expressway should be built. If one considers incremental automation of the auto in areas requiring a large number of access and egress ramps, and constructing grade separation to isolate the system from pedestrian traffic, it then becomes clear that there is actually a requirement for size and cost reduction. This might be associated with development of what has come to be known as a separate "guideway" for the automated vehicles.

Another difficult problem occurs when the system is not on a separate, single guideway. While it may be possible to detect reliably any changes in headway between vehicles following each other, it is a problem of pattern recognition of a very high order of sophistication to recognize that a vehicle, a boy with a ball, or some other object, is going to swerve into the path of the vehicle; particularly when it occurs too closely to allow for stopping or other avoidance procedures. Even detection and programming of possible avoidance procedures is yet unsolvable without a very extensive data processing capability. The first serious attempts at automating the automobile were an interesting solution to the wrong problem.

The problem concerns the city and commuters rather than intercity travelers. A system is needed to provide the safety and capacity of the interstate highway, while being compatible with the scale of the city. An equitable distribution of transportation capability to both the driver and the nondriver is also needed. The latter is particularly important when we realize that only less than half of the population is licensable to drive. The dream of the automated auto on an expressway, (many dream of it along the median strip), is somewhat like the monorail in its general attractiveness to the public, as well as to a great many professional transportation specialists.

Like the monorail, the dream partially dissolves upon further investigation, but the automated auto dream is realizable and important. The dual-mode hybrid is an evolutionary mechanism that makes staging possible, without requiring grade separation and the solution of the general pattern recognition problem first. With the dualmode concept, there is perhaps more reason to solve the commuter, the parking, and the air pollution problems within the city. Once the most pressing problems are solved, it will be relatively easy to extend automation to the intercity network, and later to convert that one lane of the expressway that has, so often, been proposed.

The Dual-Mode Innovations

The realization that a large portion of auto trips are repetitive commuter trips, came quite simultaneously to several groups of investigators. Since it has been stated that a significant percentage of the 3.5 million patents in the patent office relate to transportation, it would be impossible to list or even to collect here all the references to developments that could be categorized as the general advent of the Dual-Mode Decade, the 1970's. At M.I.T., several student design projects were initiated by the author with others starting in 1961; the first being a creative engineering class' battery powered dual-mode vehicle for the commuting trip. The results of the first study were made available to a junior design class that went on to design the "Commucar" system (3) which was also battery powered. Between these two studies, the idea of the third rail, and a power and steering control pickup arm with associated switching capability were developed. A detailed design study of a battery plus track powered system was undertaken. This design class voted to try to implement the side arm rather than the slot-car approach previously used by the creative engineering class. This Commucar project than became the predecessor of the Metran Project (4) which in 1965 investigated the evolutionary aspects of installing an automated system. The inter-disciplinary graduate systems design course that produced the Metran Report was coordinated to include some 37 students and 10 faculty members.

Among the systems implementation concepts introduced by the Metran Project were the innovation of the demand-actuated Dial-A-Bus (Genie) systems, then conceived as a feeder for dual-mode; and the personalized capsule (Perc) central city circulation system. As a result of the project, it was proposed that the development of a dual-mode system be via the transit mechanism, with the Genie system as feeder, to build up the patronage; and perhaps a Perc system of central city circulation to assist in the downtown congestion. The Metran guideway system and its precedent Commucar guideway system were conceived to be very generalized guideways that could contain and control vehicles of all types and a large range of sizes. The Glideway (5) report, a High Speed Northeast Corridor Study coordinated by the author which preceded the Metran project, specifically introduced the single car-carrying pallet for high speed intercity use. This system, as well as other freight handling pallets, were included as a part of the background of the Metran report.

The basic advantage of the pallet system might be found in the methods of implementation since they would require no modification of the vehicles, and thus could aid in the evolutionary aspects of a guideway system. Our basic position has always been that the guideway must be general enough to accommodate various vehicles, including freight and auto carrying pallets, autos, and small buses. Concurrently with the Commucar development, the Self-Transit Systems Corporation was organized in Westboro, Massachusetts, to work on what they named the StaRRcar system (6) (7). Alden's group raised some capital and began the development of a prototype vehicle that has been demonstrated on a short piece of track although it was never automated. The Alden dual-mode StaRRcar consisted of an undercarriage that operated in an enclosed box channel, and top slot supporting the vehicle by posts protruding upward through the slots.

Other approaches investigated by the Alden group included a roof-mounted system that carried vehicles on a monorail-like configuration (8). Like the Commucar project, the StaRRcar approach was to accomplish the switching on the vehicle rather than by moving a section of rail. It grasps a right-turning or a left-turning rail to switch. Since that time the Alden group has digressed in the direction of single-mode automated circulation systems and has built a breadboard prototype of a vehicle and control system. Also, simultaneous with the developments at M.I.T., the Cornell Aeronautical Laboratory's Transportation Research Department headed by Robert A. Wolf (particularly, the group under Morton I. Weinberg) began development of what was then dubbed the "Urbmobile" (9).

This system was also conceived as a battery-plus-track powered small vehicle that was to operate with a synchronous motor, using steel wheels located co-axially with the rubber tires. Throughout the various stages of development, the concepts were refined, and the most recent reports (10) discuss a chopper-driven electric motor controlled by a stable oscillator drive to give synchronous performance. It alternately rides on the steel rail or on the rubber surface changing at switches, as shown in Figure 3.1.

baumanfig3.1.jpg (62549 bytes)

The basic concept requires the wheel to drop down on its rubber surface when negotiating the switch. The rubber belts accommodate the difference in peripheral velocity of the wheels. The need for this reliance on tires negates the arguments of rubber tire reliability because the tires must certainly function at this most crucial switch location. More recently work has continued on pallet concepts similar to the Glideway pallet. While these are all perhaps just extensions of the general piggyback approach to railway utilization, a number of proposals have been made to use railway-type pallets for handling intra and intercity traffic. More recently, the McDonnell Aircraft Corporation conducted an extensive (allegedly 1 million dollars), but unpublished study of an underground pallet system for St. Louis. The McDonnell study showed that a single car-carrying pallet would cost about $10,000 per pallet and essentially pointed up some of the basic pallet features such as the vehicle's deployment problems, and loading and unloading facility requirements.

Even more recently, Professor David Wilson of M.I.T. has proposed a pallet system shown in Fig. 3.2, (11). Professor Wilson's proposal includes rack and pinion drive in order to implement the basic system synchronism. Like all the previously discussed dualmode systems, his approach also presently utilizes on-vehicle switching. One of the basic problems with pallet design is heating and air-conditioning the pallet load, as it would be undesirable to keep the automobile's engine running during the pallet portion of the trip. The loading and unloading of pallets, even if massively automated, would result in a more difficult interface between the automobile mode and the transit mode. The physical requirement would be a greater space for entrance and exit stations for efficient loading, availability of pallets at each station to accommodate peak demands.

baumanfig3.2.jpg (34012 bytes)

Dual-Mode Autos - The Guideway System

baumanfig3.3.jpg (43276 bytes)In the summer of 1968, a group of former Commucar and Metran students with a 1967 Mustang and a $20,000 grant from the Ford Motor Company began the development of a test facility to investigate some of the concepts of the side-arm power pickup system coupled with a standard American automobile. Fig. 3.3. shows the basic arm position as it is attached to a point near the firewall of the automobile. This particular dual-mode system has been named the Guideway System. The basic attachment point is chosen for a number of reasons. One, the firewall position is the stiffest lateral structure of unibody construction. Two, the arm operates near the center of gravity of the vehicle and thus enhances the stability in a skidding mode. Three, the stored arm folded into the side of the fender is in one of the least vulnerable positions with regard to the accidents that the vehicle might incur when off the guideway. Since it is quite visible to the driver, it allows for manual inspection. Such inspection would not be possible if the arm were under the vehicle or elsewhere.

One principle feature of the arm configuration is that it can operate in a power pickup channel that is open downward. This means that the channel cannot become easily filled with foreign objects, i.e. rain, snow, etc. The combined power and guard rail functions as a power pickup and steering device and as a physical barrier to keep the vehicles on the guideway, and foreign objects off the guideway. Another feature of the side arm configuration is the switching characteristic associated with letting go of the rail on one side and grabbing it with a similar arm on the other side to perform the switching function. This makes the roadway at a switch completely uncluttered. There are no track sections or grooves to be traversed by the vehicle wheels.

The basic arm system allows a mechanical fallback position that is in keeping with present automotive practice where all servo-systems have a failsafe characteristic. This is particularly demonstrated by comparison to wire-follower techniques. Another method of deducing the mechanical coupling between the vehicle and the rail results from investigation of the propulsion requirements. In order to reduce pollution, electricity is preferred because it allows the pollution to accumulate at a central power plant where it can be handled economically. The need for ease of control and reliability of the vehicle propulsion system makes the electric motor drive desirable for an automated vehicle. The use of on-board electric power obtained from a battery or fuel cell does not increase reliability in the same manner since then it would still be possible to run out of energy much like running out of gas. Significantly, fuel cells and batteries have not been thoroughly developed for this application. When they reach this stage, the dual-mode auto may well be guideway powered on the track and battery or fuel cell powered on highways. The track possibly may provide energy for recharging batteries, as recognized in early descriptions of both the Commucar and Urbmobile.

baumanfig3.4.jpg (41184 bytes)Once the decision is made for electric power pickup, then sliding contact is very quickly deduced. Putting the slider underneath the automobile results in either a protrusion or groove in the highway as a method of isolating the polarities. Side-arm configuration allows this groove to be inverted and also function as the guard rail and steering device. One of the basic problems associated with the sidearm design is the operation at stations where access to the vehicle is required. Fig. 3.4 shows the automobile in a station with the arm flexed downward into a surface level groove. A collection channel below this groove protects it from being destroyed by obstruction of foreign articles.

baumanfig3.5.jpg (26268 bytes)

Figure 3.5 shows the schematic design of the basic side-arm follower that has been selected a s the first prototype. The arm configuration allows power pick up from two angular plates, one on either side. The experimental configuration utilized an angle of 120 to keep the follower system captive in the guide rail, thus isolating the follower dynamics from the vehicle dynamics. The follower arm must have three degrees of freedom; roll, pitch, and yaw. Roll and pitch are both accommodated by pivots shown in Fig. 3.6. The yaw degree of freedom is indicated by the exursion of springs on the power pickup pads. These springs are mechanically limited so that if the vehicle pulls heavily to one side, (such as would be the case in tire failure or wheel failure modes,) the arm simply accommodates the extra force by the mechanical stop. The resulting slider location in this maximum deflection position still allows the opposite slider to be in physical contact with its power conductor. The presently designed pad springs are constructed from a stack of preformed spring-washers that give a constant-force spring characteristic to the slider.

baumanfig3.6.jpg (72075 bytes)

The Mustang system was installed on an abandoned railroad right-of-way with a 200 foot section of track. The track is basically folded up out of 3/8 inch steel plate with a composition board insulator and a 120 included angle inside power conductor also formed from steel plate. Fig. 3.7 shows the automobile on the prototype guideway. To conduct power tests, a series-wound DC motor (the starter from a diesel truck) was mounted to the rear of the differential with a chain drive and over-riding clutch For minimum cost configuration, the power supply of the prototype was comprised of a set of 6 volt batteries with SCR switching to turn the power on or off in 6 volt steps. This could have been installed trackside, but would have necessitated bringing out power lines for recharging. Rather, the batteries and power control system were installed in the trunk of the Mustang with power led out through the arm, across a shunt, between the two poles of the power rails, and back into the vehicle in series with the battery supply and motor.

baumanfig3.7.jpg (71932 bytes)

This configuration allows for testing of the complete track, the power-pickup system, and some investigation of electrical noise characteristics, as well as experiments with the basic feasibility of side-arm steering. Two steering experiments were undertaken. The first was the so-called "stiff arm mode" in which the spring loaded side arm forced the vehicle from side to side, both with the steering wheel locked and in a free condition. From this operation it was found that it is possible to steer through the small angles required by a simple forcing system, utilizing the natural caster of the tires. In later experiments, the side arm was mechanically coupled to the steering linkage of the automobile via direct hydraulic lines from the side arm cylinder to an equivalent cylinder on the steering link. This configuration caused the front wheels to steer as a direct function of the distance of the vehicle from the track. The vehicle and track were operated in all kinds of roadway conditions, including glare ice, deep snow, and normal cinder track configuration. In the stiff-arm mode, extreme acceleration on glare ice allowed some skidding from which the vehicle could not recover if it had gone through too great an angle. The mechanical coupling overcame this difficulty. On an actual guideway there would likely be guide rails on either side of the vehicle. Thus, sidewise skidding of the rear and extreme angles of skidding would be impossible. In addition to these considerations, an actual guideway would stay hot enough with use to keep snow or ice from accumulating.

An automated system might have programmed snow plows and ice melters which perform when required. The mechanical steering system would be the preferred mode, with fallback to the passive mode of non-crucial failure when necessary. A preliminary computer analysis of the vehicle dynamics of standard autos with side arm steering indicated no cause for concern at any feasible guideway speeds.

Since the author is accepting a new position at Carnegie-Mellon University, it is anticipated that the vehicle and guideway will be transported to the Pittsburgh area, where it is expected that further testing of this and similar systems will continue.

The Future of Dual Mode Activities

Transportation in America has always been at the apex of focus between the public and private sector. The private autos on public roads, private aircraft on public airports and airways, private trains on private rails, and all possible combinations in between, have been one of the mainstays of our transportation heritage. The dual-mode auto can be envisaged as either a transit system or a highway system and, as such, further confuses the issue as to who is primarily responsible for its development. Since the automobile is such a large sector of the basic transportation economy and since transportation itself is something on the order of 20% of the GNP, it is difficult to convince the public and thus Congress that the major development should not be done by the automobile interests. On the other hand, the automobile interests point out that the government has subsidized the development of the most recent transportation thrust in aviation, and that they believe it is government's role to finance this next thrust. As one auto executive put it, "Anyone who develops an automated vehicle system will need us. And when they are ready to pay for it, we will be ready to build it".

Systems as large as would be required for implementation of dual-mode, exceed the scale at which any patent protection on componentry would be significant. Few other mechanisms exist for providing a private organization the incentive to embark upon so great a venture. There is some interest in such a venture by aerospace corporations because, as they point out, they would like to get into a field where profitability is determined by the establishment of a consumer price rather than a fixed percentage established by government contracts. The present economic recession may significantly dampen interest in the private sector of the economy, but it is completely clear that dual-mode activities will and are continuing in a number of places. There is little doubt that the most significant urban and mass transit benefits would ensue from the development of this system. Furthermore, the entire electronics and aerospace technology that is presently in search of a new field could fruitfully engage in the design and testing of these systems which would be a significant and substantial improvement to urban transportation.


1. Hamilton, William F. II, and Nance, Dana K, "Systems Analysis of Urban Transportation", Scientific American, Vol. 22 1, pp. 19-27, July, 1969.

2. Gardens, K:, "Automatic Car Controls for Electronic Highways", General Motors Research Lab., General Motors Corp., Warren, Mich. Rept. GMR276, June, 1960.

3. Baumann, Dwight M., Blanco, Ernesto E., and Mann, Robert W., "Commucar - An Intra-Urban Transportation System", 1966 National Transportation Symposium; ASME; New York, May, 1966.

4. Project Metran: An Integrated, Evolutionary Transportation System for Urban Areas, Interdepartmental Student Project in Systems Engineering at the Massachusetts Institute of Technology; M.I.T. Press; Cambridge, Mass.; 1966.

5. The Glideway System: A High Speed Ground Transportation System in the Northeast Corridor of the United States. M.I.T. Report No. 6, M.I.T. Press, Cambridge, Mass.

6. Descriptive Literature from Alden Self Transit Inc., Bedford, Mass.

7. Brush, S.G. et al Patent No. 3,363,584, Jan. 1968.

8. Alden, W.L., Patent No. 3,2 54,608, June 1966.

9. Gilmore, C.P., "How You'll 'Drive' the Amazing Urbmobile", Popular Science, Oct. 1967.

10. Robert A. Hayman, et al., "Bi-Modal Urban Transportation System Study", Vol. 1, Report PB 1 78 286, Cornell Aeronautical Laboratories, Buffalo, N.Y., prepared for U.S. Department of Housing and Urban Development, March 1968.

11. Wilson, D.G., Bisbee, E.J., Clarkeson, and Paul I,  "Quadramode Transport: A Class of Controlled Systems", Urban Engineering and Transportation, presented at the ASME Winter Annual Meeting, Dec. 1968.

home2.gif (1492 bytes)

Last modified: June 21, 2003