by Peter Benjamin

Volpe Transportation Systems Center, U.S. Department of Transportation , Cambridge, MA.

This paper is a summary of a four-volume report entitled Analysis of Dualmode Systems in an Urban Area (DOT-TSC-OST-73-16A-1), published in 1973. All of the documents are available from the National Technical Information Service, Springfield, VA, 22151.  This paper was originally published in the Proceedings from a conference entitled Personal Rapid Transit II: Progress, Problems, Potential, University of Minnesota, December, 1973, pp 95-106. [Updates shown in red text]

The automobile provides convenient, flexible, relatively low-cost transportation, and is thus the overwhelming choice of urban travelers. Currently, increasing concern is being voiced over relatively uncontrolled growth of automobile travel. Noise and air pollution, the divisive effect of ribbons of concrete cutting through neighborhoods and the plight of minority groups and the poor traditionally displaced by new urban freeways are being recognized. Consequently, public pressure has developed in opposition to new highway construction. In fact, in a number of urban areas, new roadway construction has come to a virtual standstill.

While demand for transportation continues to grow, conventional transit systems have been unable to attract significant ridership or provide the service desired by travelers. What is needed is a transportation system with the apparent advantages of the automobile but without the associated congestion, pollution or large right-of-way requirements.

Dualmode transportation systems have been suggested as alternative transportation forms with the potential to meet this need. A dualmode vehicle is one which travels under manual control on the street network for some portion of its trip, and operates under automatic control on an exclusive guideway for some other portion. Thus low density collection/distribution functions could be accommodated at low capital cost using existing street facilities, while high-density routes with common origins and destinations for many travelers could be automated.

Automation provides the potential for: 1) achieving increased guideway capacity through close-headway operation, 2) allowing safe high-speed travel without congestion, and 3) providing increased free or productive time to drivers by relieving them of their duties. Electrically powered dualmode vehicles may help to reduce air pollution, and guideway design may permit minimization of noise transmission to adjoining areas. Dualmode transportation systems have the potential to provide door-to-door transportation equivalent to the automobile in convenience, and thereby may attract ridership from highways and reduce the problems of congestion.

The U.S. Department of Transportation, through the Systems Analysis Division of the Transportation Systems Center has conducted an economic feasibility analysis of dualmode transportation systems in order to provide information upon which to base research and development priority decisions.

The analysis was conducted in a 1990 Boston scenario, in which an extensive dualmode system was presumed to exist. The scenario was chosen only to provide meaningful base data. The study was not a proposal for a dualmode system in Boston or a transportation plan for that area. As a basis for comparison a 1990 transportation plan for Boston projected by the Eastern Massachusetts Regional Planning Project was also analyzed. The analysis was oriented toward examining urban-wide applications of the dualmode concept, as opposed to limited-service systems for specific purposes.

For the purposes of the analysis, performance levels were specified with the assumption that the appropriate technologies (such as command and control) would be developed sufficiently to permit their attainment. Continued technological development is required to achieve these capabilities.


The proposals for dualmode systems that have been made by various institutions, companies, and developers were examined for their basic technological and application elements. These basic elements were categorized and grouped according to common characteristics, from which evolved generic baselines which represent classes of proposals. Each of the baselines discussed herein consists of a combination of personal vehicles and buses, thereby providing the user with alternative choices in using the system.

The three baselines examined in this report in detail are described in Figure 1. The pallet system consists of conventional private automobiles which are driven onto pallets that operate on the guideway, and 20-passenger buses which do not use pallets but interface directly with the guideway. [For a current example of a pallet (or car ferry) system, being developed, see the MegaRail website]

A typical trip on the system is depicted in Figure 2 [updated with RUF vehicles by Palle Jensen]. A standard automobile is driven manually on the street network to the guideway entrance. The car is then driven onto the pallet which operates automatic control on the guideway. If the destination is outside of the urban core, the car is driven off the pallet at the appropriate exit and then manually operated on the streets to the destination. The already extremely congested downtown area of Boston does not rationally invite discharging large numbers of dualmode vehicles onto the streets. Consequently, no exit from the system is permitted in this area. Upon arriving at an urban-core station, the automobiles are unloaded from the pallets and are parked in garages, with no street access permitted. The riders then walk to their destinations or transfer at the garage to the existing local transit system for downtown collection and distribution.

The buses were assumed to operate on fixed routes and schedules on the streets in the suburbs. Upon entry to the guideway, the driver leaves the vehicle and it operates automatically. In the downtown area, transfer to transit it required.

The automated-highway baseline consists of: 1) automobiles in their automated mode, interfaced directly with guideway, and 2) buses which operate in the same fashion as the pallet-system buses. 0ff the guideway the autos perform in the conventional manual mode. [For a current example being developed, see the 9-minute Qwiklane video]

The new small-vehicle baseline consists of: 1) innovative small personal vehicles specifically designed for operation, and 2) dualmode 12-passenger minibuses. In this case a dense guideway network with stations easily accessible by walking is provided in the central business district, thereby eliminating the downtown transfer to transit. The small personal vehicle is designed for individual use, but is owned by and rented from the system. [For current examples being developed, see the Personal Rapid Transit and Dualmode overview webpages]

As shown in Figure 2, [updated with RUF vehicles, by Palle Jensen] for suburban collection the vehicle is driven manually to the closest station, and if the destination is in the downtown area the user leaves the vehicle upon arrival at the station closest to the intended destination. For a return trip, a vehicle, but not necessarily the same one driven in, is provided at whichever downtown station is chosen. Thus a vehicle is always guaranteed at a downtown station, but no permanent correlation between particular individuals and vehicles exists. The minibus operates as a dial-a-ride vehicle in the suburbs.  [An excellent current example is the Danish RUF dualmode system]


A single dualmode guideway network was designed for all baselines, with some adjustments required to meet the peculiarities of specific systems, particularly in the downtown area. In keeping with the objective of minimizing community disruption, an attempt was made to use existing rights-of-way whenever possible.

The dualmode guideway network designed for the Boston scenario is depicted in Figure 3. Most of the stations, shown as circles, provide for entry to and exit from the system. Stations in and near the central core are indicated as three different types. The squares are stations where dualmode users interface directly with existing Boston rapid transit systems. Exit from and entry to the street network is provided, and parking at the station is encouraged by reduced parking costs. The stations indicated by the diamonds permit only parking and a transit interchange, with no street access. To discourage parking in areas of higher land cost, higher parking rates were established at these stations. The one station shown with a triangle has no transit interface, but is within walking distance of the main financial district. The new small-vehicle system substituted a dense network in the downtown area as shown in Figure 4.


Intuitive analyses of various systems indicated that for urban-wide applications, mixed vehicle fleets are more attractive than either personal dualmode vehicles or dualmode bus systems alone. In all cases examined, mixed fleets provided greater ridership and higher revenues as well as lower cost per passenger trip than single vehicle systems. The bus systems provide service for the "transportation poor" while the personal vehicles maximize the diversion of travelers from the highway onto the guideway. The mixed fleet not only provides a logical implementation sequence, starting with the bus and then adding the more complex operations of personal vehicles; but also, provides the flexibility to meet changing patterns of transportation demand.


The service levels achieved by the three dualmode baselines and by the l990 plan are compared in Figure 5. All of the dualmode systems attained more than a 10% modal split--more than the transit split in the case. Since less than half of the trips in the region are of sufficient length or are located so that they are candidate dualmode trips, the 16% split attained by the new small-vehicle baseline actually represents the attraction of more than 30% of the potential dualmode users. Dualmode attracts as much as 53% of the peak-hour downtown-bound travelers. Although the transit and dualmode systems were designed to be complementary, considerable competitive characteristics remained. Therefore, the dualmode systems might be expected to fare even better in other scenarios without the large existing investment in transit.

Figure 5  Service Comparisons
  1990 Plan Pallet Automated Highway New Small Vehicle
Route Miles 114+29 (1) 249 249 261
DM Modal Split   12% 11% 16%
Transit Modal Split 10% 6% + 2%(2) 6% + 1% (2) 6%
Peak Period Surface Arterial Speed (mph) 15.8 17.7 17.5 18.6
Average DM Trip Speed (mph)   24 23.8 25.3
Typical DM Trip Time Savings (min.)   17 16 19
Daily Regional Time Savings (years)   19.4 19.3 36

Notes: (1) A+B = "A"  new highway miles + "B" new transit miles   (2) A% +B% = "A" transit riders + "B" riders transferred from duamode at parking garages

The diversion of riders from the highways onto the dualmode guideway reduced highway traffic congestion, as evidenced by an increased peak-period surface arterial speed. The dualmode systems themselves achieved as much as a 57% increase in door-to-door travel speed compared with a similar trip under the 1990 plan. For a typical dualmode trip the average user saved in the vicinity of 15 to 20 minutes compared with the trip he would have taken if the 1990 plan had been adopted. The combination of reduced highway congestion and generally higher average speed of dualmode travel resulted in the saving of as much as 36 years of travel time every day.


The total capital cost of the urban-wide implementation in the Boston scenario of the various dualmode baselines ranged from 1.6 billion dollars to more than four billion dollars, as shown in Figure 6. Smaller systems could be constructed at proportionately lower costs. Just over one billion dollars (in the form of 114 miles (71 km) of highway and 29 miles (18 km) of rapid transit extension) of proposed construction in the 1990 plan would not be built if dualmode were installed. This, in effect, represents a capital cost savings incurred by the adoption of a dualmode system.

Figure 6  Comparison of Costs and Impacts
  1990 Plan Pallet Automated Highway New Small Vehicle
System Capital Cost ($ M) 1,020 (1) 2,360 1,620 4,200
System Vehicle Capital Cost ($M)   682 (2) 83 (3) 1,970
Annual DM Capital and Operating Cost ($M)   724 625 972
DM Cost per Passenger Trip ($) (4)   1.83 1.81 1.90
Regional Annual Trans. Pollutants (M pounds) 309 343 1,032 336
Daytime Noise Impacts (Households) (5) 41,000 431 1,032 336
Household Displacements (5) 58,000 6,100 6,100 5,400

Notes: (1) Highway and transit construction if dualmode is not built; (2) Pallets and buses; (3) Buses only; (4) Annualized capital plus operating costs; (5) Associated with new transportation facilities

The purchase price of 420,000 system-owned small personal vehicles accounts for nearly half the capital cost of new small-vehicles. In this baseline a significant benefit accrues to the large number of users who rent a dualmode vehicle and forego the purchase of a second family car. For the automated highway, the system vehicle capital costs include only the purchase of buses, since the personal vehicles are privately owned and operated. Both pallets and buses are included in the vehicle cost for the pallet system.

Capital costs were annualized by applying a 10% interest rate and depreciating each individual element of each system over its lifetime. Total annual dualmode capital and operating cost reflects total door-to-door transportation costs to society, including such items as the operating and depreciation costs of private vehicles during the off-guideway portions of pallet or automated highway dualmode trips. They range from a low of just over 600 million dollars to almost one billion dollars, with the variation largely representing differences in service levels and ridership. The annual costs per passenger trip for all systems were approximately the same.

Figure 7 shows how the profits and losses of the system operator would change with variations of up to 25% from the nominal assumed fare levels. Revenues of all systems can equal or exceed operating costs, but local capital subsidies would be required to meet total system costs.


Figure 6 also compares some of the regional impacts of the various dualmode alternatives and of the 1990 plan. The pollution output associated with the generation of electrical power for the pallet and new small-vehicle baselines was included in the figures, which are for all transportation modes in the region. The high power requirements of the pallet alternative caused by the necessity to move heavy pallets and the vehicles on them, as well as empty pallets, resulted in the pollution levels of this baseline exceeding the 1990 plan. Longer trip lengths, high speed operation, and diversion from transit caused the automated-highway vehicles, powered by internal combustion engines, to contribute to a greater total pollution level than the l990 plan. The only significant reductions in pollution levels were achieved by the new small vehicles.

Most of the route miles of the systems analyzed in this scenario were accommodated on existing rights of way or were tunneled. Dualmode systems, with twice the route mileage of planned highway and transit additions, displaced only 10% of the number of families which the 1990 plan of construction would have moved. Although the pallet and automated-highway vehicle baselines had fewer route miles than the new small-vehicle baseline, the former required large parking garages for the storage of private vehicles, thereby causing somewhat greater displacements. Largely because of guideway structure and location dualmode systems caused only 1% of the noise impacts associated with the 1990 plan. Thus, at a time when community pressures are making acquisition of new right of way for transportation systems increasingly difficult, dualmode systems can significantly reduce neighborhood disruption and division as well as avoid the displacement of minority groups and low income families--the traditional victims of new transportation system construction.

Cost and Benefits

The annual costs and benefits of the dualmode alternatives relative to the 1990 plan are summarized in Figure 8. Costs and benefits were calculated on a regional basis for all transportation modes. Thus the costs include dualmode operating costs and annual capital debt service, differential cost savings due to reduced highway and transit construction and maintenance, and changes in the costs incurred by individual motorists operating their vehicles. The benefits costed include travel time savings, relocation savings, accident savings, changes in pollution costs, and changes in land values and tax revenues. They do not include benefits from such items as decreased neighborhood intrusion, additional job accessibility, or regional economic stimulation, which either could not be adequately quantified, or were considered un-costable. The 1990 plan was used as a base, and all costs incurred beyond that level and all benefits obtained above that base were included.

Figure 8 shows that the regional benefits of the mixed-vehicle dualmode systems, for urban-wide applications, are more than twice the costs. Of the systems analyzed, the new small-vehicle system has the greatest total benefits, the largest net benefits, and the highest ratio of benefits to costs. Moreover, this baseline attracts the highest level of ridership, and revenues exceed the operating costs by the greatest increment. It obtains this, however, at the highest capital cost of any baseline and thus requires the biggest capital subsidy. These capital costs, although quite large, are not inconsistent with the costs of any large urban transportation system, and on a route-mile basis are lower than those of the New York Second Avenue Subway and the projected unit costs of the Washington, D.C. (METRO), Baltimore, and Atlanta rapid transit systems.

Because of the generally conservative assumptions used in this analysis, the results presented tend, except where noted, to project a conservative case for the dualmode systems. As parametric analyses indicated, more optimistic projections would improve the general picture, although the relative results (differences between alternatives) would not be expected to change significantly.


Urban-wide applications of dualmode transportation systems were considered in this analysis. However, such systems will not come into existence instantaneously; rather they will grow over a period of years. Because the system characteristics will change during this implementation period, various dualmode alternatives will provide the greatest effectiveness at each stage.

Initial implementation of dualmode will, in all probability, occur in a high-demand density corridor. The limited extent of the guideway and restricted number of origins and destinations would tend to discourage the purchase or rental of personal dualmode vehicles. Thus, if personal vehicles are to be used in a limited-scale system, the pallet would appear most appropriate alternative.

A limited corridor with relatively high demand occurring between common origins and destinations permits the attraction of sufficient ridership and the achievement of high enough load factors to make bus systems very effective. The low vehicle capital and operating costs per passenger mile of bus systems with high load factors in such limited applications provide the opportunity to install a working dualmode system at minimal cost and still achieve acceptable ridership.

Because dualmode buses can be operated on guideways at relatively large headways compared to those required to move the same volume of people in personal vehicles, bus systems would appear ideal for the initial developmental stages of these systems, when command and control technology is in its early stages of maturity. Of the bus systems examined, the minibus with dial-a-ride service achieved the greatest ridership, and would appear to be most suitable for an initial dualmode application. Personal vehicles could be introduced at a later date to increase utilization of the guideway, with the pallet being the most attractive candidate so long as a limited diversity of origins and destinations is available.

This analysis suggests that as the network expands, the new small-vehicle alternative, in scenarios similar to Boston, provides the greatest benefit/cost ratio. If sufficient funds are not available for implementation of this alternative, continued expansion of the pallet or introduction of the personal automated-highway vehicle may be preferred, despite their slightly lower benefit/cost relationship. Were the new small-vehicle system to be implemented, however, continued use of pallets or pallet-like vehicles for freight would be desired to further diversify system usage.

The introduction, at a later stage, of a coordinated off-guideway non-dualmode dial-a-ride bus feeder and on-guideway personal vehicle (PRT) system has the potential to further increase ridership while decreasing costs per passenger trip through increased bus load factors. The evolution of an urban-wide application of dualmode transportation, therefore, would utilize a number of the concepts examined during its various stages of growth.


Although this analysis was conducted in the Boston scenario, these conclusions are considered to apply to many other large urban areas as well:

  • dualmode systems appear to be sufficiently attractive to warrant further technological development.
  • For urban-wide applications, dualmode systems which include mixtures of buses and personal vehicles are more effective than either fleet alone
  • A dualmode transportation system benefits from the use of various dualmode concepts throughout its development
  • For the first stage of implementation of a dualmode network, the dualmode minibus seems most effective. Capability should be provided for subsequent expansion to the use of personal vehicles and buses

The results presented here are for only one scenario and for urban-wide implementation of the concept. It is expected that different demand patterns, population densities, and urban forms would lead to some differences from the results obtained in this analysis. The rather extensive rail rapid transit system in existence in Boston, together with the extremely dense central business district having poor surface arterial circulation, forced the network design and operating policies for the alternatives examined to be considerably different from those which would be expected in other scenarios. The prohibition or dualmode vehicles from the downtown streets would probably not be necessary in cities with a lower population density and better central business district arterial circulation. The differentially priced downtown and peripheral parking with transfer to transit might not be as desirable elsewhere, nor would the extensive tunneling for the downtown network necessarily be required. In Boston the analysis examined dual mode and transit as complementary, but also competing modes. This would not be the case in cities which do not have an existing investment in rapid transit.

An overview of current dualmode development projects and issues is provided by a Dualmode overview webpage and a Dualmode Debate webpage. A major Dualmode Study performed at the Texas A&M University by the Center for Environment, Energy and Innovative Transportation (CEETI) is also available on-line. A link to it is provided at the Dualmode overview webpage.

Last modified: September 27, 2009