Personal Rapid Transit Works in Simulation Only - An Answer to Professor J. Edward Anderson
by Vukan R.Vuchic
Note: This is a response to an article by Prof. Anderson that was published simultaneously, in the December 20th issue of the Urban Transportation Monitor
Before discussing imaginary ubiquitous networks of PRT guideways over city streets, or operations of thousands of vehicles at extremely short headways, it is useful to clarify some basic concepts about transportation systems: vehicle sizes, guidance vs. highway driving, and area coverage.
For low density travel, such as in suburban residential area, small personal vehicles are operationally and economically most efficient. As travel density increases, vehicle capacity logically should be increased not so much to "amortize drivers wages," as Prof. Anderson claims, but because larger vehicles do not need extremely short headways (which reduce reliability of operation), and because of sheer economies of scale: hundreds of minibuses in Manila are much less efficient on heavily traveled streets than articulated buses in Curitiba; transporting 400 persons across the country is more efficient in a jumbo jet than in 30 Cessnas. Coal is transported in 100-ton, not 20-ton, rail cars not simply to minimize "wage amortization"!
The dominant reason that thousands of people travel in cars on freeways instead of using buses or trains is that they have dispersed origins and destinations which transit cannot easily serve. The main advantage for use of PRT instead of AGT (its closest relative in the family of modes) would be if PRT could go to many more destinations. But this would be the case only in very elaborate and expensive networks: even then, PRT cannot compete with private cars, because guided systems inherently have far fewer destinations than highway vehicles. Thus, the PRT concept attempts to copy the ubiquity advantages of private cars and efficiency of rail transit due to exclusive guideways, but it actually cannot come close to either one of these competitors.
There is a fascination with short headways PRT could operate. Theoreticians like to analyze "subsecond headways" which, even if technically feasible, would neither be desirable nor achievable for many vehicles in sequence. Off-line stations would, presumably, allow undisturbed passage of mainline traffic while vehicle alighting/boarding is performed. But if, for example, a train in Chicago unloads 75 persons at the PRT station, they will need some 40 PRT vehicles, and that boarding would take a much longer time than boarding two AGT or one light rail vehicle. Capacity of boarding can be increased by building long platforms, but large structures make stations even less acceptable in many urban settings. Simulation may show that station operations are fast, but that depends on how realistic are the assumptions used in the simulation model.
Prof. Anderson has always claimed that PRT is competitive with other modes because its vehicles will be very light and simple, and the guideway inexpensive compared to other modes. The development of the system by Raytheon does not confirm this: the prototype vehicle cost and weight per seat (500 kg) are higher than on most transit modes. The guideway is very complicated. (some photos of the Raytheon PRT 2000 test facility are available).
The common gap between theoreticians and practitioners is particularly great with respect to PRT systems: the PRT concept attracts the interest of some vehicle designers and control engineers, and a number of operations researchers find challenges in optimizing individual operational processes. However, this concept has never found the support of persons involved in urban transportation planning, or in the design and operation of real world transit or taxi systems. Anderson's argument that the interest of some public officials and Raytheon's investment in R&D of this concept prove the system's feasibility is shaky at best. Let us remember the Monocab PRT, then its magnetically suspended version developed by Rohr, German Cabinenbahn and French SAFEGE: these and dozens of other "new systems" were developed at great cost and found no applications.
During the 1970s there were serious proposals for development of " dual-mode " systems with vehicles running on guideways or driven on streets. When GM and others attempted to design such guideways and ramps for a specific city, most of the "system assumptions" had to be modified so much that it became clear the system would not be feasible. Then, millions of dollars were spent on designing and simulating "GRT" or Group Rapid Transit 12-passenger guided vehicles running at "subsecond headways"a concept even more unrealistic than PRT. Where are GRT systems now?
Similar to monorails -- which even today attract the fascination of laymen and the general public but are inferior to rail transit (with exception of a very few specialized applications) -- PRT may look attractive as a "system of the future." The chances that this system can compete with cars in suburbs, rail in major corridors, and AGT or buses elsewhere, are small indeed. Consequently, as long as practical operational and economic evaluations are considered in selecting urban transportation systems, PRT will remain a "system of the future" only.
Vukan R. Vuchic is Professor of Civil Engineering at the University of Pennsylvania. This article appeared originally in the Urban Transportation Monitor of December 20, 1996. It is a response to an article by Prof. Anderson published in the same issue. It is posted here with the permission of the publisher, Dan Rathbone.
Last modified: October 21, 2008