Return to home page

Personal Public Transport

Martin V. Lowson
Advanced Transport Group
University of Bristol

Accepted for publication in Proceedings of the Institution of Civil Engineers, Transport 1999

Introduction

Present transport systems, especially those based on the car, have been extraordinarily successful. But this very success has led to considerable difficulties. To quote from the Foresight Panel Report on Transport (1995)

" Most people treasure the enrichment that personal mobility brings to their lives
- most are increasingly concerned about the problems caused by everyone else's car ownership "

There are massive rush hour jams in all major towns and cities. The London Journey Times Survey 1996 shows that straight line speeds from start to destination in the central area are 2.8 mph by car and 2.6 mph by bus.

Transport is responsible for around 25% of all energy used in the UK, and more than half of all the airborne emissions (Transport Statistics Great Britain 1996). The car has now killed more than 25 million people world wide. On the other hand almost 30% of people in the UK still do not have direct access to a car, while present cities require up to 30% of their area to be given over to roads and car parks. Further, the success of the car has disadvantaged those who are unable to use it directly, such as the young, old or disabled. Provision of public services which can meet their requirements has demanded significant subsidies. It is projected that car traffic will increase by more than 50% by 2025 (National Road Traffic Forecasts 1997). The very success of the car has led to a compelling problem requiring urgent resolution.

Nevertheless, despite the manifest inefficiencies, and the resulting environmental problems, the preferred choice for travel is most often by personal, as opposed to public, transport. 80% of all trips over a mile are by car, and 90% of all vehicles on the road are cars. Around 85% of current household expenditure on transport and travel is on motoring. (Transport Statistics Great Britain 1996). This testifies to the very considerable value which individuals place on the ability to convey themselves from place to place at their own choosing.

It is clear why an individual will choose a car in preference to current public transport. The car provides a fully flexible system which can respond immediately to meet personal transport wishes. This need appears to be fundamental, and provides a central rationale for design of a transport scheme for the next century.

Current public transport systems were conceived, designed, and built at a time when the dominant element of any city was the principal city centre, which contained both the major shopping areas and the major work areas. Conventional public transport such as trains or buses, which are corridor based and collective, served such cities well. In the early development of cities most residential areas clustered around the principal public transport corridors, as can be seen on old maps.

Single centre cities are now a thing of the past, and the types of public transport systems which served them are outmoded. Cities are now multi centre. The great majority of new shopping areas, business parks, industrial estates, supermarkets, hospitals, etc., have been constructed away from the city centre. Since all of these are usually far from residential areas, the new pattern of the multi centre city generates a new type of strongly decentralised "anywhere to anywhere" travel demand. This is because the new forms of city have developed around the capability offered by the car. The new demands cannot be met effectively by the old forms of collective-corridor public transport, which match old forms of cities. Public transport will not meet present transport needs unless it is radically changed to match the form of modern cities.

During the past 200 years the principal forms of surface transport have moved from canal to rail to car-road, and become more oriented towards personal travel. From the wider historical perspective it is clear that a new form of surface transport will come into use during the next century. The only question is whether this new form appears sooner or later. It seems that it is reasonable to expect it to appear soon, due to society's growing appreciation of the problems of current transport modes.

Kuhn (1970) pointed out that science and technology do not develop seamlessly, but rather by major and painful changes in accepted wisdom. Such changes are preceded by a period of profound uncertainty as the accepted solutions become plainly unsatisfactory at solving the problems. The critical nature of the problems of present day transport is reflected in the present political debate. There is now a general recognition of deep rooted difficulties faced by present approaches to transport and the need for major change. According to a recent survey 92% of people believe that the most important problem requiring research today is low pollution transport.

Great effort has been expended in devising ways in which current solutions can be improved. The results have been, at best, marginal. Under such circumstances, history tells us that success is more likely to come from innovative than traditional approaches.

Pucher and Lefèvre (1996) say (p201) "Urban decentralisation has greatly increased travel distances and has reduced the importance of trips to and from the city centre, which public transport serves best. Travel between and within suburbs, on the other hand, is growing fast in all American, Canadian and European Cities, and it is precisely this sort of trip pattern for which the car is a virtual necessity. Such suburban trips are usually too long for walking or cycling, and they do not generate high enough travel volumes in route corridors to make (current) public transport economically feasible. Thus suburbanisation has sharply reinforced the trend towards ever greater use of and dependency on cars". As stated in the recent Government Consultation Paper "people use the car because they are denied real choice". They are denied real choice because the city has developed into a form which cannot be served effectively by existing types of public transport.

The difficulties faced by current collective public transport systems are fundamental. Further, as Pucher and Lefèvre point out (p203) "Huge subsidies have been injected into public transport in most countries, but those funds have not succeeded in producing high quality public transport networks, .... accessibility by public transport has not improved over the years in spite of huge investments and subsidies. Public transport policies have failed to create a satisfactory alternative to the automobile."

The objective of the work of the Advanced Transport Group work has been to define the principal features of a public transport system for the next century, meeting future needs for flexible personal transport, while being highly acceptable in an urban environment. The detailed aims of the project are to provide an urban transport system which can:

reduce energy use and emissions by a factor of 10 over cars
halve urban journey times
provide transport for all, 24 hours a day

That is to provide a transport system which can meet both the users desires for effective, convenient, and flexible transport and the non-users need for minimal impact. We have also sought ways to meet this goal within the confines of current technology.

2. Historical Analysis

Figure 1 shows the development of surface transport in the UK over a quarter millennium, in terms of the build rate, non-dimensionalised against the maximum for each system. The data used to develop this graph is based on official statistics for rail and road, (Transport Statistics Great Britain 1996, Mitchell, 1975) and an analysis of parliamentary bills for canals kindly provided by Mr R.Jamieson, Archivist at the National Waterways Museum, Gloucester.

These changes in the nature of transport have been driven by cost. Only in comparatively recent years have environmental considerations become a major determinant in transport planning. Figure 1 shows a continuing process of change in the nature of transport systems, which have also become more oriented towards carriage of people. In fifty years time it is reasonable to assume that another transport system will be in place. Some important clues to possible characteristics of this new system come from further historical analysis. A fuller discussion is given in Lowson (1998).

The move from canal to rail involved a change of both vehicle and infrastructure. The same is true of the change from rail to car-road. Thus for a new system it seems reasonable to look again to a change of both vehicle and infrastructure. In our studies we have found that a simultaneous change of both vehicle and infrastructure provides exciting new opportunities to solve current problems.

The key technological discovery for the train, Richard Trevithick's high pressure boiler, was invented just after the major peak in canal building. At the same time the railway track was developed, so that rail supplanted canals rather quickly. The key invention for the car-road system, the internal combustion engine, was invented just after the second major peak in railway building. Unfortunately Daimler did not define the infrastructure required, and it was not until old roads and cart tracks were converted to the modern trunk road and motorway system that the car-road system became a dominant force in transport.

We are now past the major peak of motorway building. The historical analysis suggests that the key technologies which will drive the next form of surface transport may well be available now. The task is to identify these, so that we do not rely on variations of surface transport systems originally invented in the last century to answer the problems of the next.

3 An Analysis of Transport Problems

The conventional division into public and private transport is not very helpful. A more useful classification is shown in Figure 2, which analyses the transport as either personal or collective and as full area, anywhere to anywhere, or corridor. Figure 2 also assigns current forms of transport under this classification. Most current forms of transport fall into either to the top left or the bottom right hand boxes.

All present public transport can be classified as collective corridor systems, while all private transport is classified as personal full area. This provides the starting point for an explanation of why it is that current public transport only satisfies 20% of present trip demand.

As already noted, corridor systems are effective for journeys in and out of the city centre, but they provide limited access to the suburbs unless both starting and ending destinations happen to lie on a corridor. In addition any form of collective transport suffers from two major disadvantages. First, the larger scale inevitably means that there will be waiting. Second, it is impossible to obtain simultaneous speed and accessibility.

This latter point is illustrated in figure 3, which shows an estimate of achieved trip speeds by transport which has various maximum speeds, but is required to start and stop at regular intervals. This analysis follows that presented by Hamilton and Nance (1969). A (high) acceleration of 0.25g has been assumed. The figure shows that high maximum speed capability is of no value if stops are close together. The bus, with stops typically 250m apart, gives good accessibility, at least along the corridor, but the frequent stops provide a fundamental limitation on speed. If stops are further apart, say 5 km as in trains, then speed between stations can increase, but there is little accessibility. Typically passengers will get to the train station by car. Thus current forms of collective corridor public transport suffer from fundamental limitations.

But a form of public transport which was in the top left box of figure 2, which was both personal and full area would not suffer these problems. The Advanced Transport Group has sought to define such a system.

The fundamental need for anywhere to anywhere transport is reinforced by statistics on the trip demands. Data on projected trip demands by car between pairs of zones in Bristol during the peak hour was kindly provided to us by Bristol City Council. This projection is based on real traffic census data. An analysis is shown in Figure 4. The figures shown are all for the number of trips projected between zone pairs during the peak hour. There is no reason to suppose that these figures would not be representative of other British and European cities. Thus Figure 4 confirms that the true demand in a modern city is anywhere to anywhere. It would be almost impossible to devise a conventional collective transport system to meet these very diverse needs.

The data also shows that actual demand is surprisingly low, far lower than everyday experience would suggest. For Bristol the average demand over the whole of the built up area is 600 trips per square km over the peak hour. This equates over a 100m by 100m block of only one trip demand every ten minutes during the peak hour. This explains why it is that public transport companies have so much difficulty providing a viable service over the whole city. The congestion which is experienced in all cities is a consequence of funnelling a modest demand into limited corridors, with stops at many of the crossing points.

Thus it can be concluded that a new approach is required, one which is public and also personal, which can respond effectively to demand, and which does not require intermediate stops. Clearly this will demand a new network and vehicles dedicated to this requirement. This is consistent with the argument for a new infrastructure and vehicle which has been previously identified on historical grounds. The Advanced Transport Group has completed a study which does define such a system.

4. A Possible System

4.1 Basic Synthesis

The key features of the new transport system generally follow logically from the basic objectives set and are defined below.

Personal Transit: The key distinction to be made in the present system, following the discussion in Section 3 above, is to meet personal transport needs. Thus the vehicles must be sized to be effective for transport of individuals. However the possibility of families or small groups travelling together must not be excluded. This implies a maximum capacity of about four people.

Full Area In the same way it is clear from the previous discussion that the system must provide as far as possible an anywhere to anywhere capability. The network should, in principle, cover most of the city.

Off Line Stations All stations must be off line so that stopping at one station does not impede the free passage of other vehicles.

Separated Track: For the system to be effective it is necessary that the infrastructure, i.e. track, is dedicated to the new system, in the same way as for a new train or tram system. Sharing an existing infrastructure would force all the inefficiencies of the existing system (road/rail etc) into the new system. This does not exclude common use of rights of way, but track and hence vehicle separation is essential.

Elevated Track: If the track is dedicated to the system, then separation from the existing system, and the maintenance of existing land uses by pedestrians etc forces much of the track to be carried overhead. Thus effective design of the overhead track structure is a crucial element of the system. Wherever possible track would be at ground level to reduce cost, particularly at stations, but in an existing city opportunities for ground level tracks are limited. An alternative is underground, but the considerable additional expense of underground construction will normally rule this out, except in very particular circumstances.

Automated: Automation is also a fundamental requirement if vehicles on the system are to be shared by a number of successive users during the day. This demands automatic recirculation of the empty vehicles. Automation is also well matched to the needs of a flexible personal service, and has crucial cost saving implications over a manned public system. Automation to full modern standards of safety also has significant potential benefits over human control for safety.

One Way Tracks: On a fully automated service the key requirement is to be able to provide a vehicle on demand. Providing the network is sufficiently flexible, the precise direction of initial travel is of little consequence in overall journey time, except for the shortest journeys. (This can be seen from Fig 6) However, a restriction to one-way tracks reduces track scale, cost, and intrusion significantly. It also very considerably simplifies design of interchanges and stations, which are liable to be a major, probably dominant, cost element of the infrastructure. This does not exclude two way tracks where suited to particular situations.

In Vehicle Steering: In principle, change of direction on the track can be accomplished either by train like "points" or by a vehicle based system. Choice of an in-vehicle system means that the track is an entirely passive structure, thus giving a useful cost reduction for the track. In any case there are overwhelming advantages of in-vehicle steering, since it is not feasible to switch points at the short headways that must be anticipated in a personal transit system.

Note: the conclusions to this point are entirely consistent with the arguments of those who have supported various forms of Personal Rapid Transit (PRT) system over many years. A further analysis of previous work is given in Section 5 of the present paper. Some divergence from the traditional PRT arguments emerges in the points to be made below.

Dual Mode: A truly flexible personal transport system should offer a capability of movement from home to work. Many individuals are unwilling, or unable, to get to a station to pick up a centrally controlled vehicle. Thus to offer effective competition with the car, it is believed to be essential to provide the opportunity for starting and finishing the journey at home. This requires a "dual mode" approach in which vehicles which are personally controlled off the track can operate on the track under system control concurrently with those controlled entirely centrally. Such a capability need not be available in the system from the start, but it is crucial that this facility can be made available in later versions of the system.

Rubber Tired Vehicle and Matching Track: Providing an opportunity for dual mode use forces a major feature of the vehicle design. The vehicle which leaves the track will have to possess largely traditional car forms of wheel and steering. This forces matching requirements on the track, and also on the vehicles which never leave the track.

U Form Track: The simplest form of track guidance is wheels running about a vertical axis against a track guidance rail, already in operation in kerb guided buses. An alternative possibility is guidance using a buried cable, already a developed technology. This would have special benefits in circumstances where the separation between the present system and traditional forms of transport was incomplete. However the present system has been configured around simple mechanical guidance. A U-form track is necessary to provide simple switching. Inverted T tracks such as used in monorails require special mechanisms to provide wheel clearance at the switch points, and are therefore impractical for a network with many elements. Running in a U track should also provide useful psychological reassurance to the passengers, and reduce the visual impact of the passing vehicles.

Minimalist Infrastructure Design: Previous studies of equivalent systems have demonstrated that the two most critical issues are the cost and visual impact of the infrastructure. A minimalist approach to infrastructure design offers the best hope of controlling these difficulties. It is likely that this implies a steel structure.

The proposal may be summarised as an automatic personal taxi system running on its own dedicated guideway network. Figure 5 shows a computer image of the system which has emerged. It has been called ULTRA, an acronym for Urban Light TRAnsport. Vehicle weight has been reduced to a minimum in order to reduce resources required throughout the system. Vehicle weight is currently projected at about 1000kg (1 tonne) all up. This is very significantly smaller than current public transport vehicles.

Detail vehicle issues are not regarded as a fundamental to the system. The key issues are automation and small scale. Several of the design features were also determined by the wish to provide for a dual mode capability which could be introduced within the system at a later date. Conceptually, either internal combustion or electric power could be used for the vehicles, and both have been evaluated. For the core system an electrically powered vehicle, with trackside power pick-up has been selected. Studies to date suggest that there are few, if any, issues with the IC option, which has technology very parallel to that in a small car. Indeed a variety of fully automatic IC vehicles have already been tested for other applications. There will be further detail practical issues to be faced with the electric option, for example, snow and ice. Such issues are regarded as an important part of the next, project definition, stage of the work.

The system would be used by buying a ticket at a ticket machine. This would cause the central control to program the vehicle to travel automatically to the destination by the most effective route. Empty vehicles recirculate automatically on the network for the next fare. Thus the system is both personal and public.

4.2 Form of Network

Figure 6 shows idealised forms of networks for use in a PRT system. The most obvious is a grid alignment as shown in Figure 6a. The key disadvantage of this is that each junction has to be at two levels thus leading to problems in both cost and visual intrusion. An alternative system is shown in Figure 6b, based on a series of loops. This has the advantage that overhead structures are avoided at intersections, but results in less efficient use of the guideway, since parts of the guideway have to serve both N-S and E-W traffic. Any system installed in an actual city must conform to the topology of the city, so that none of the idealised solutions in Fig 6 can be used in practice without significant modification. However, examination of issues in Bristol has made it clear that 6b is a useful basic model for a practical system.

Virtually all forms of current transport systems are multi-speed with low speed collector areas and high speed sections for longer distance elements. The provision of a two speed system for a new approach is therefore an important potential opportunity. There are already major linear routes in all towns, such as motorways, major urban trunk roads, railways and rivers. Using these rights of way for an additional form of transport would result in relatively small impact on the community, both in terms of added severance and of visual intrusion. Since the scale of the system is small, typically 2 metres across or less, there is every prospect of being able to find parts of the rights of way (e.g. disused railway lines, verges) which can be readily utilised for the system, still maintaining separation from the other transport mode. It also seems feasible on the major linear routes to run the system in two directions. The concept of two-way high speed routes at grade level is attractive, since costs are reduced significantly. These high speed routes could be associated with slower feeder loops which would be one-way, perhaps elevated more frequently. Stations located on the feeders would be of smaller size because the lower speeds required in that part of the system reduce the length of the acceleration/deceleration lanes. Note that this is again an idealised view. Any real system would have to meet a variety of practical issues of detailed routing.

Present evaluation suggests that stations should be about half a kilometre apart to minimise walking. This implies that guideways will also be about 0.5 km apart. This is a sparse guideway network, and compares to about twelve conventional roads per km in a typical urban setting. There should be a station within about 250m of most locations. Reduction of the number of stations per unit area only provides a square root increase in average distance travelled to reach the station at start or destination. Because of this, recent studies have suggested that a cost effective starting point is a somewhat sparser system.

A section of the guideway proposed is shown in figure 7. Figure 7a gives a typical ground level layout, and Figure 7b an example overhead design. These designs are due to A Kerr of Ove Arup. It can be seen that the objective of a minimalist infrastructure has been satisfied. The guideway is of small scale, only two metres across. This means that it can be built at low cost, and with minimum intrusion into the city. As already discussed, the track is passive providing a further gain in both cost and resource usage. Figure 7b shows an overhead design which is based around standard steel I beams. This provided an excellent basis for cost estimation. Further detail design studies of the overhead structure suggest that it is possible to reduce the structure depth to 50cm for a span of 20m. This has significant potential benefits in terms of both cost and visual intrusion. A fuller description of the work on planning and related issues is given by Medus and Lowson (1999).

4.3 Energy Use

A major driver for the present work has been the desire to achieve substantial improvements in transport energy use and environmental impact. Analysis, demonstrated that a remarkably large proportion of the fuel usage, and therefore emissions, of present cars was parasitic, being required to overcome engine friction. This is shown in Figure 8, based on a model originally presented by Ross (1992), but adjusted to account for the smaller engine displacement volume (1.6 litres) and smaller weight of UK cars.

However, during the peak hour average speeds are now less than 20kph, and energy usage at these low speeds in conventional cars is further compromised, both due to engine drag terms and because of additional losses due to braking and idling during the stop-start process. Typical results for the car are shown in Table 1. These are based on the Ross model for the case of four one minute stops every km, and show that less than 15% of the energy expended under these conditions is used usefully in propelling the vehicle. But, in addition commuting car journeys have a significant penalty from cold running. Waters (1992) gives data showing that the loss for the typical 5km journey is as much as 160%. However the most recent studies of cars with modern fuel control systems, Sérié and Joumard (1997) suggest far lower cold start penalties, typically around 10%. Cold start penalties for emissions are generally rather higher, but vary significantly for different pollutants. The figures for energy usage for the proposed vehicle are based on a smaller weight and electric power, with a basic thermodynamic (0.38) and transmission (0.87) efficiency taken equal to that for the petrol engine. It can be seen that, comparing traffic during peak periods, a basic gain of over a factor of ten is projected. Set against this gain is the energy required to move the empty vehicles. Nevertheless the projected energy savings compared to the current car-road system during the peak period are very substantial. Equivalent emission savings can be anticipated since these are directly related to the reduction of fuel used.

Table 2 gives a summary of estimates by several workers on energy use in terms of MJ /passenger km by various forms of transport. Generally, there is reasonable agreement between the various figures. The major discrepancy is between the two figures for light rail. The Coffey and Lowson (1996) data was based on actual data obtained from the Manchester and Tyne and Wear metros. An interesting trend is the improvement of car energy consumption with time, primarily reflecting improvements in engine technology. The second trend is the higher recent estimates of energy use by the bus. These reflect current statistics on the use of smaller buses which have been found to be more effective for transportation purposes, although less energy efficient. However, the essential conclusion from Table 2 is that there is surprisingly little difference in energy use between current public and private modes of transport once full account is taken of the actual levels of utilisation.

Over the whole day, an energy utilisation of about 0.55 MJ/pax km is projected for the proposed system, ie a benefit of a factor approaching 4 over cars and better than 2 over current public transport. As already noted, the benefit increases very substantially for a comparison of use during peak periods

For inner city use it is preferable to use an electrically powered vehicle since this will eliminate emissions within the city. An analysis was undertaken of battery powered vehicles. This demonstrated that batteries were inadequate for transport use and likely to remain so. To put this into perspective it would take more than ten tons of conventional batteries to equate to the energy that can be stored in a fuel tank. Much of the power usage of battery powered cars is required simply to move the mass of the batteries. Further, the chemicals used in batteries, particularly proposed advanced batteries, are far from environmentally advantageous. It was concluded that exclusive battery power was impractical.

Since in the present system the vehicles are confined to a guideway there is the opportunity to provide the power directly via a pick-up. This is an attractive option since it removes the requirement for substantial energy storage on the vehicle, although batteries are required for back-up. Current designs of the vehicle use a proprietary conductive power pick-up. Serious consideration should also be given to inductive pick-ups, but present analysis suggests that this technology is not quite ready for mass exploitation

4.4 Control

The automatic control of these systems is an obvious issue. Extensive studies suggest that there is no major problem of principle. There have been many previous studies of control of equivalent systems (see discussion in Section 5) and experience from these studies is available to guide decisions on the detail of the system to be used here. Nevertheless extensive simulations have been undertaken of the control of a complete network. This includes responding to variable trip demand and the inclusion of empty vehicle management algorithms. These have confirmed that there are no problems of a fundamental nature.

In all cases the issues are parallel to those in queuing theory. While demand levels are low it is straightforward to provide service. However, as demand levels approach maximum vehicle availability queues can build very quickly. It is planned in the present system that any such queues will be at the station rather than on the network. There are a variety of simple strategies which can achieve this objective.

However the simulations have also revealed that optimising the stations to meet the demand is a far more complex task than first thought. The problem lies in the highly variable nature of the demand. Stations have to be designed in such a way that there is both a taxi rank of empty vehicles for new outgoing demands, and a rank of empty berths to meet incoming demands. Statistical variability makes providing an adequately high service level more difficult than anticipated. This is the subject of current study. There is little problem with station design in areas of low usage. It appears to be more worth while to replicate stations in areas of high demand than to attempt to design overcomplex platforms. This is an issue which can only be optimised once more detailed cost information is available.

An analytic model for the station filling statistics can be developed. This only examines the requirements for vehicle movement events, not any issues associated with timing. Suppose the probability of arrival of a passenger is a and the probability of departure is d, and the station is formed from B berths from which passengers can alight and/or board. The probability of r berths being filled is governed by the table below

where f is an arbitrary constant. It can be confirmed that the solution given remains valid after a further arrival or departure, and thus by recursion is valid for all cases.

The constant f is found by the requirement that the sum of all probabilities for all possibilities is 1. Hence

f= 1/ SUM [a^(B-1) + a^(B-2)*d + a^(B-3)*d2 + ..... a^(B-r)*d^r ...... + a*d^(B-2)+ d^(B-1)]

The solution also admits an arbitrary constant, but this must be zero to satisfy the result for a=0, d=1 when all berths have zero vehicles apart from the first.

If the station has no vehicles then a vehicle must be drawn from store. This is described as a "pull". Equally if a vehicle arrives when all berths are full then an empty vehicle must be despatched. This is described as a "push ". The results above show that

the probability of a push is f*a^B
and the probability of a pull is f*d^B

The consequences of this are shown in Figure 9. The results are symmetrical in arrival and departures so that the number of pushes for a given arrival rate is equal to the number of pulls for the same departure rate.

When the number of arrivals is equal to the number of departures, (a=d=0.5) the probability of an arrival generating a push or of a departure causing a pull is 0.5/B i.e. inversely proportional to the number of berths. This demonstrates one of the design issues in the present automatic system. Even if there are, for example, three berths at a station, and arrivals balance departures, the statistics of the arrival-departure process mean that one third of all demands will require an empty vehicle movement. Thus the station design is a critical issue in operational effectiveness. In fact the design is further complicated by timing issues, but the above statistics capture much of the major design problem.

4.5 Other Aspects

Many other aspects of the system design have been examined, and it is not possible to provide detail of all of these in the present paper.

Safety is a critical matter. An initial study was reported by Lowson and Medus (1997). In broad terms travel by car results in about ten times more risk of death or serious injury than travel by present public transport. However more than 90% of car accidents are caused by driver error. Thus an automatic system which eliminated the driver would offer immediate prospects of a major improvement in safety. Further, in the proposed system all vehicles travel in the same direction on the guideway. Active safety devices which warn of potential hazard will be a necessary feature of the design. Consistent with other objectives, a target safety level has been set of a factor of ten improvement over cars, i.e. broadly similar to current public transport.

There are now many automatically controlled devices which interact with the public, and safety standards have evolved to a point where the engineer can design to achieve exceptional safety performance. Automatic landing systems in aircraft are an outstanding example. A survey has been performed of all automatically controlled public transport systems. Not one serious incident has been identified which has arisen from a failure of an automatic device. Thus it is believed that the safety issues can be addressed. However , it is the perceived safety of automatically controlled systems which is the critical issue in the acceptability of the proposed system. Safety will remain one of the major issues throughout the detail system design.

An analysis has also been made of the potential use by people with disabilities. Because of the nature of the system it is readily accessible to the young and the old. It should also be noted that some forms of disability such as blindness, epilepsy, or lack of motor skills which would preclude use of the car do not preclude use of the new system. Travel to the stations is a negative factor for some disabled people, but the dual mode version of the vehicle opens up the system to the disabled in exactly the same way as the motor car. Analysis in a report by Lowson and Davey (1996) suggests that the system would be available to around 75% of those with disabilities compared with 45% who can take advantage of present cars.

Two issues of obvious relevance are vandalism and personal security. It is suggested that the basic approach to crime and vandalism prevention is via management, supported by good design and careful choice of materials. A good certainty of detection via video cameras etc is a significant deterrent. Entry to the vehicle is only possible with a valid ticket. This reduces the risk of vandalism. Damage within the vehicle would probably be detected by the sensor suite. On detection the vehicle would be diverted to a manned station where the issues could be dealt with. On personal security issues it should be noted that compared to present systems there is little or no waiting. No-one is required to travel with another passenger. In the event of personal problems during travel, emergency buttons in the vehicle would permit an immediate stop at the next station. There would also be voice communication between each vehicle, stations and central control.

4.6 Planning Issues

Perhaps the most critical aspect of the whole project is the acceptability of the system in a typical city. Many issues come together under this general heading. The first is the realistic possibility of introducing the system in an existing city with all of the limitations that this implies. Although the system is only 2m wide a real guideway has to negotiate all of the existing city structure of roads, pedestrian zones, conservation areas etc. By its nature any city has limited unexploited space. Also, some of that currently taken up could eventually be set aside for use by the new system. Thus the location of stations can be an issue of particular significance, particularly since the stations also require deceleration and acceleration lanes which extend the land take required. The ability to provide adequately large radius of curvature on the track to provide transit without undue speed reductions is a further issue.

A critical issue for the cost predictions is the balance between ground level and overhead structures. Any ground level guideway will automatically create a barrier between one side of the track and the other. This will lead to issues in community separation. In principle the small scale of the system means that overpasses need only be around 2m high, but these may not prove acceptable. Alternatively the guideway may have to pass over the road. This is a more demanding requirement since current legislation requires that the clearance is 5.7m. With a standard 6% gradient this implies 100m transitions. It may be that a better solution in some cases would be to tunnel, but this option has not yet been studied in depth.

Currently many communities separate out parts of the road for bus lanes. In principle this option could be available here. However this raises new planning issues on which there are at present no guidelines.

Perhaps the most critical issue is visual intrusion. In order to avoid the problems of separation it is inevitable that significant parts of the system will be overhead. This eliminates the problems of separation but introduces the new problem of visual intrusion. The basic system design has already addressed this problem by putting considerable emphasis on a minimalist infrastructure. As already noted, present design have been able to reduce the depth of the overhead sections to 0.5m. This is considerably less than existing road structures. Indeed the structure is less intrusive than typical pedestrian walkways. It is believed that the structure as currently defined should be broadly acceptable visually. However this has not been tested in any formal way with either public or planning authorities.

All of these issues are minimised if parts of the city which are already used for transport are reused for the new system. Thus the use of old railway lines or the verges of main roads provides not only a low cost ground based solution, but also one which can significantly reduce planning problems.

However there is also important planning gain from the new system. It will complement the car where it is introduced, and thus minimise or, if it can be made sufficiently attractive, eliminate the growth in car use. This could result in major benefits as some parts of the city currently used for roads or car parks can be returned to use by people. A network over a significant area of the city could also provide a major benefit to other modes. It can act as a feeder for the wider use of trains, would be very effective in increasing the attraction of park and ride facilities, and provide a good match to an increased level of rural bus service.

4.7 Costs

The cost of the system is a critical determinant of the overall practicability of the system. It is not appropriate to give detailed cost figures in the present paper, but initial estimates have been favourable. Preliminary estimates of the cost and viability are inevitably prone to errors in assumption, but the figures should provide an initial basis for evaluation. Cost figures have been based either on professional advice or actual budgetary quotations. The costs shown in Table 3 are based on estimates originally made by A. Kerr of Ove Arup in association with the design studies underlying Figures 7. The overhead cost is based on an estimate for 10m column spacing 5.7m high. The estimates include an allowance for both diversion of utilities, and for the provision of electrical supply. These estimates were provided by SWEB. The cost of a single lane of conventional urban road is also shown. For comparison it may be noted that the maximum capacity of the two systems in terms of vehicles per hour is essentially equivalent at about 1800 vehicles per hour per lane.

Estimates for the relative amounts of at-grade, elevated, and underground track required for various situations have been made by Medus (1997), following detailed examination of sites in Bristol. Further results can be found in Lowson and Medus (1999). It was found that most areas in Bristol permitted some routings without excessive difficulties, either in terms of cost or planning issues. As the track density required in a given area increases then the difficulty of finding low cost routes increases. The analysis showed that introduction of the system at about 1.5 km/km2 would allow a total cost per km of £400k, (ie £600k /km2) whereas a dense network of 4 km/km2 would cost £650/km (ie £2600/km2). Including initial estimates for stations, control, maintenance etc suggested that the network would be viable even at the higher density at a fare of about £1 per vehicle for a 5km trip, ie a fare comparable with current buses. The figures above show that cost increases rapidly with guideway density.

Estimates of the potential modal split between the new system and existing cars and buses have been made, Lowson (1997). These were based on standard logit cost of travel models, calibrated by the known split between bus and car for trips to the Bristol city centre. The results suggested that between 25% and 40% of current car journeys would transfer, and an even higher proportion of bus trips, even at fares of in excess of £1 per trip.

Initial studies suggest that potential passenger demand remains strong at the lower densities. Thus introduction of the system at a relatively low density, and therefore cost, appears very practicable. There appears to be a satisfactory relation between the projected costs and income potential for the system. The figures indicate a capability of meeting a minimum return on capital employed of over 10%. Thus although the figures are preliminary, they do provide a basis for encouragement.

It is not claimed that the system will replace the car, certainly not in the near term. As already noted, the new system will complement the car where it is introduced. Our analysis does suggest that the ability to serve a complete network, without waiting, mean that the system would be considerably more attractive than current public transport.

5. Comparison with Previous Systems

The ideas discussed here are generically known as Personal Rapid Transit, or PRT, and have been examined by others. The first paper which clearly outlines a system which is essentially PRT is that by Fichter (1964). An extensive study was undertaken in the US at the Aerospace Corporation starting in the 1960s. The basic philosophy developed from this work was presented by Hamilton and Nance (1969) and considerable detail is provided in the book by Irving (1978). The ideas were further developed by Anderson (1977).

In the UK a proposal put forward by Blake (1967) culminated in the Cabtrack system which was extensively studied in the late 1960s; Langdon (1971). The principal problems in Cabtrack were the cost and visual intrusion of the infrastructure, which was over 3m in depth. Cabtrack also used multi-level exchanges which increased significantly both the cost and visual intrusion of the exchanges. This was exacerbated by the Cabtrack control policy which required acceleration/waiting/deceleration lanes at every junction. Generally Cabtrack appears to have concentrated on the vehicle issues, and given inadequate emphasis to the problems of the infrastructure, including stations. Compared to Cabtrack, the present system has identified many areas for reduction of cost and visual intrusion.

At that time there were also several other studies. In addition to the US study mentioned, other systems include Aramis in France, recently the subject of a book Latour (1996), CabinenTaxi in Germany, and CVS in Japan. None of these reached fruition. It is interesting that many of these systems went through the same metamorphosis, moving from personal rapid transit to group rapid transit. Unfortunately any form of group rapid transit brings with it the fundamental difficulties of collective travel that were discussed in section 2. Automated group rapid transit with small vehicles raises the further problem of a possible enforced ride with undesirable travelling companions. This issue would be a fundamental barrier to acceptance today.

The reason for this metamorphosis can be traced to the selection of the incorrect application for the system. The present system can be highly effective for travel over the whole of modest scale cities such as Bristol, or in the outer areas of very large cities such as London. The system as currently configured is not well matched to the problems of the centres of the largest cities, where conventional solutions such as the Underground are reasonably effective. Unfortunately Cabtrack, Aramis, and Aerospace, took respectively London, Paris, and Los Angeles as their principal target cities. This significantly distorted the conclusions, and led to minitram type approaches, which discard the key conceptual advantage of the approach. In the case of Aramis the designers were always committed to a corridor system, and so excluded one of the major transport advantages of the approach from the start.

One system with many technical features in parallel with the present ideas is that at Morgantown in the US. This has fully automatic cabs on a dedicated guideway. It has been running successfully for 20 years and has now carried 50 million passengers without incident. Although called PRT, the cab capacity is 21, and it is in reality a collective-corridor system linking two parts of a University campus. Nevertheless, it does demonstrate that a fully automatic demand responsive system can be technically successful.

Currently the "PRT 2000" system is being developed in the US at a cost of $50M. It is hoped that it will open for passenger operation in Rosemont, a suburb of Chicago, in 2002. The PRT 2000 system is conceptually very parallel to the ideas presented here, and has many of its advantages. The system was originally based on the ideas of Anderson (1977) but has undergone a significant increase in size. Compared to the system described in the present paper the US system thus has significant disadvantages both in cost, and in particular in visual intrusion, which will be of particular importance in a European context.

Generally it seems that PRT studies undertaken in the 1960s and 70s were ideas which were ahead of their time. We have been able to make significant improvements over these systems, and eliminated a number of the technical weaknesses of old solutions, especially in the approach to infrastructure. Also, today we are able to exploit a wide range of new technologies, especially computing. The development of PRT 2000 for Chicago demonstrates that these ideas are now coming of age.

Conclusions

It has been shown that existing technologies brought together in a new system can offer a new approach for transport which is both public and personal, is well matched to the multi-centre form of modern cities, and is truly sustainable.

The system can provide significant benefits for the user and the non-user alike. Compared to current collective public transport, the system is projected to offer a benefit of around three in trip time, while also providing travel anywhere to anywhere on the network at any time without waiting. Compared to private cars at peak periods the system has a factor of ten reduction in total emissions, together with zero emissions in the city. It can be available to all, including the young, the old, and the disabled. The system offers the prospect of significant (factor of ten) improvements in safety, and dramatic reduction in resource use which reflect in significantly lower costs of construction. The system can also enhance the attractiveness of other modes, especially train and park and ride. Initial analysis shows that the system can be viable at fares comparable to current collective public transport

Most importantly, by reducing the area of the city given over to roads and car parks, the new system provides the opportunity to return cities to use by people. The system proposed offers significant gains in energy/ emissions, effectiveness and resource usage and can provide effective and sustainable personal mobility for the 21st century.

Acknowledgements

The work reported here has been undertaken by the Advanced Transport Group at the University of Bristol with support from EPSRC, the Rees Jeffreys Road Fund, and the Department of Transport. Inputs from all members of the ATG team are warmly appreciated. Note: Fuller details of the work of the Group can be found at http://atg.fen.bris.ac.uk

References

ACEC 1976 "Passenger transport: short and medium term considerations", Advisory Council on Energy Conservation Paper 2. Dept of Energy HMSO
Anderson, J.E., 1977 "Transit Systems Theory" Lexington Books Mass.
Blake, L.R., 1967 "A public-transport system using four passenger self routing cars" Proc Inst Mech Eng Vol 181 pt3G pp132-144
Coffey, R., and Lowson, M.V., 1996., "A comparative analysis of energy consumption and emission of urban transport systems", in "Urban Transport and the Environment II", Eds. Baldasano, Recio and Sucharov, Computational Mechanics Publications
Fichter, D., 1964 "Individualised automatic transit and the city " B.H Sykes Chicago
Foresight Panel Report on Transport 1995 HMSO
Government Statistical Service 1996 "Transport Statistics Great Britain". HMSO
Government Statistical Service 1997 "Journey Times Survey 1996 Inner and Central London" The Stationery Office
Hamilton, W.F., and Nance, D.K., 1969 "Systems analysis of urban transportation" Scientific American Vol 221 No1 pp19-27
Hammarstrom, U., 1988 "Equivalent energy consumption and exhaust emissions" Swedish road and traffic research Institute (VTI) Paper 567 Linkoping
Hillman, M., and Walley, A., 1983 "Energy and personal travel: obstacles to conservation" Report 611 Policy Studies Institute
Hirst, E., 1973 "Energy intensiveness of transportation" Transportation Engineering Journal; Proc Amer Soc Civil Eng., Feb 1973 pp 111-122
Irving 1978 "Fundamentals of personal rapid transit" Lexington Books Mass.
Kuhn, T.S., 1970 "The Structure of Scientific Revolutions" University of Chicago Press
Langdon 1971 "The Cabtrack urban transport system" Traffic Eng and Control Vol 12 pp634-638
Latour, B., 1996 (Translation: C. Porter) "Aramis or the love of technology" Harvard UP
Leach,G., Lewis,C., Romig,F., van Buren,A., and Foley,G., 1979 A low energy strategy for the United Kingdom. Int Inst for Envir and Dev. Science Reviews Ltd London
Lowson M.V. and Medus, C.E., 1997 "Initial Safety Considerations for an Advanced Transport System", chapter in "Safe Systems", Eds. Anderson and Redmill, Springer Verlag
Lowson M.V., 1997 "Combined study of Modal Split and Economics for an ULTRA System" ATG Working Paper 97-9 December 1997
Lowson, M.V. 1998 "Surface Transport History in the UK - Analysis and Projections" Proc Instn Civ Engrs Transp Vol 129 Feb pp14-19
Martin,D.J., and Shock, R.A.W., 1989 Energy use and efficiency in UK transport up to the year 2010 Report for Dept of Energy by Chief Scientists Group, Energy Technology Support group Harwell
Medus, C.E., 1997 "Planning Personal Rapid Transit Networks: A Preliminary Evaluation" ATG Working Paper 97-4 University of Bristol Sep 1997
Medus, C.E., and Lowson,M.V., "Planning PRT Networks: The Case of ULTRA" Paper at ASCE APM Conference Copenhagen Jun 1999
Mitchell, B.R., 1975 "European Historical Statistics 1750-1970" London UP pp581-584
North, B.H. 1993, "A Review of People Mover Systems and their Potential Role in Cities", in Proc. Instn Civ. Engrs Transp. 100, May, p95-100.
Pucher, J., and Lefèvre, C., 1996 "The Urban Transport Crisis in Europe and North America". Macmillan
Ross, M., 1994 "Automobile fuel consumption and emissions: Effects of vehicle and driving characteristics" Annu Rev Energy Environ. Vol 19 pp 75-112
Sérié E., and Joumard, R, (1997) "Modelling of cold start emissions for road vehicles" INRETS Report LEN 9706 Feb 1997
Waters, M.H.L., 1992 "Road vehicle fuel economy" TRL state-of-the-art Review 3 HMSO
Wood, C., 1995 " Energy use by different passenger modes" Transplan Norwich (quoted in Potter,S., "Vital travel statistics" Landor Publishing 1997)

Return to home page