The (Il)logic Behind PRT

Post by Steven Dale

The Morgantown PRT. Those vehicles sure do look personal don't they? Image by Brian M. Powell.

Personal Rapid Transit (PRT) is a technology that’s been dreamed about in transit circles for roughly the last 60 years with little to no progress. For those unfamiliar, the concept of PRT is based around the following 4 principals:

  • Small, automated vehicles with seating for 2 – 8 people.
  • Vehicles available on demand at stations throughout the system.
  • Direct from origin to destination. Vehicles do not have to stop at intermediary stations.
  • Non-linear networks of stops, thereby eliminating the need for transfers.

The concept is that for public transit to compete with the private automobile, it needs to replicate the comfort of the car. Fair enough.

The transit-geek-gadgetbahn-aficionado in me would of course love to see PRT sometime in my lifetime. But that’s based upon blind hope and little else. The fundamental logic behind PRT just doesn’t work. Here’s five reasons why:


ONE. Vehicle Capacity. The appeal of Personal Rapid Transit is that it’s personal.

And yet if every vehicle were loaded with only one single rider, there would be plenty of wasted capacity and seats. As a PRT system typically has only one single guideway, the system would basically just be replicating a single lane of under-capacity cars. There is, however, a solution to this problem. Which leads me to my second problem:


TWO. Solving the vehicle capacity problem negates the whole concept behind PRT.

The only way to solve the problem outlined in the previous point is to enforce a ‘carpool’ mentality. How popular is carpooling? Station attendants would necessarily have to force riders to ‘buddy-up’. 8 person vehicles would be filled by 8 people whether they were traveling together or not. Suddenly it’s not personal. Suddenly you have 8 different people traveling to 8 different locations. Which leads me to the third problem:


THREE. Station Attendants will cost money.

Any cost savings that PRT imagines would be erased by the need for station attendents to enforce carpooling during peak hours. Despite having these station attendants, it’s unlikely that the attendant will be able to group passengers according to their destinations. As such, we have a fourth problem:


FOUR. People going in 8 different directions must travel in 8 different directions.

So now it’s rush hour and we’ve got 8 different people traveling to 8 different destinations. Now the algorithm used to control the vehicles must calculate a linear route that stops at each destination sequentially. And that would be utter insanity. Imagine if you and your fellow rider were traveling to destinations at the exact opposite ends of your respective city!

You could solve this problem by giving everyone their own vehicle, but to do so leads us straight back to point ONE. The only real way to deal with this issue is to institute fixed routes, which leads me to the fifth problem:


FIVE. The Appeal of PRT is the Elimination of Fixed Routes.

If suddenly every PRT system is a linear fixed route, then what we have is nothing more than an Automated People Mover that has the ability to skip stations. Note, however, that as 8 different people with (presumably) 8 different destinations are using this souped-up APM, riders will still be faced with the situation of stopping at stations different from their destination.

It shouldn’t surprise anyone that the most famous “PRT” system in the world is the Morgantown PRT in Virginia -which shares a surprising resemblance to the situation I’ve just described.


SIX (BONUS!). Google’s already invented PRT.

It works and is a driverless car.

Want more? Purchase Cable Car Confidential: The Essential Guide to Cable Cars, Urban Gondolas & Cable Propelled Transit and start learning about the world's fastest growing transportation technologies.

Related Posts Plugin for WordPress, Blogger...

Want more? Purchase Cable Car Confidential: The Essential Guide to Cable Cars, Urban Gondolas & Cable Propelled Transit and start learning about the world's fastest growing transportation technologies.


  1. Ah, yes. I remember the old german telefone boxes. ;) Proof: http://upload.wikimedia.org/wikipedia/commons/5/5b/Telefonzelle_in_Bochum_01.jpg
  2. Actual experience with the Masdar City PRT system is that vehicle occupancy averages 2.0 on weekdays and 2.5 on weekends - double what you have assumed. For high demand situations such as Indian cities, we have developed methodologies that facilitate ride sharing in very large systems with no station attendants while allowing those who want to pay extra to pay per vehicle, rather than per ride, and maintain the personal experience. Please educate yourself about the realities of PRT by visiting our website www.prtconsulting.com where you will find pictures, videos, papers, links, etc.
  3. Okay, here I've got to respectfully disagree with you on a few things. While as a fellow 'gadgetbahn enthusiast', I agree that PRT is a nifty idea never properly implemented, and may never be in my lifetime, there's a few points here that I feel are straw men. Firstly, point number 1: unused capacity. Most PRT usage estimates, if they are being genuine, use the estimated load of about 1.2 persons per vehicle when estimating capacity. That's why most systems deserving of the name have a per vehicle capacity between 2 and 4 persons. Secondly point number 2: 8 persons per vehicle is Not PRT. Morgantown's misnomer notwithstanding, the proper name for that type of proposal is Group Rapid Transit, which is a similar, but different beast. As your remaining points derive from these fallacies, they come across as a 'straw man', sorry to say. I think that Gondola transit is a sound and implementable idea, with a proven track record. But I'd be much more inclined to listen to objections to PRT if they were more firmly grounded in an understanding of the technology. To the general public, there honestly doesn't appear to be much of a difference between PRT and gondolas. And from a technical standpoint they face similar problems and solutions. For example, you've recently been posting about stacked and doubled up systems. You're not the only one: http://openprtspecs.blogspot.com/2010/12/capacity.html I present all this in the spirit of informed debate. I'm a fan of both concepts.
  4. Steven , this system only presently runs at about a six second headway because there is presently no demand for shorter headways. However, a number of modern PRT systems have been certified for three second headways. Cabintaxi proved it could operate at 1/2 second headways but did not prove safety or reliability at those headways. I have no doubt that PRT systems will be operating at one second headways in the reasonable future. One second headways and an occupancy of 4 results in a respectable theoretical maximum guideway capacity around 14,000pphpd. Remember, PRT's strength is network capacity, not guideway capacity.
  5. Steve, I'm glad that at least now we are on to more technically accurate arguments. Cabinentaxi, a PRT system developed to the point of extensive prototyping in Germany in the '70's, demonstrated 2.5 second headways with 'brick wall' stops. It also demonstrated that 0.5 second headways were possible, but could not meet the defined standards for railway stop safety (as in it was possible, but the standard called for the ability to 'brick wall' stop one vehicle in the separation distance). http://faculty.washington.edu/jbs/itrans/cabin.htm Note also that it was designed for over-under track usage. It was also designed with GRT capacity available for peak use times, presumably to operate on scheduled service. The system was never built beyond the prototype stage due to a change in government, and thus spending priorities. 2.5 seconds, though, should be a reasonable goal for system headways, which would make a peak capacity of 1728 persons per hour per track possible. Assuming a trunk line configuration of 4 tracks sharing one set of towers, 2 tracks per direction would yield 3456 persons per hour. Given computer managed routing, there's no reason you couldn't run 3 tracks one way and one the other during peak times, yielding 5184 persons per hour in the peak direction. Another factor to consider is that PRT systems typically assume a more spread out network wit multiple paths to the same destination. I'm not a particularly strident supporter of PRT. I'm just interested in the idea enough to have done a bit of homework. There are plenty of problems with existing proposals and implementations. The currently understood model of PRT is pretty much defined by the problematic systems at Masdar and Heathrow, and it will probably be mired that way for some time. PRT is always an 'almost there' technology, because there are always several stages of prototyping in the way. I think that Gondola systems are much more achievable in the near term, and may very well be cheaper in the long run. The transit geek in me is just as interested in a trip to Morgantown to try their (misnamed) PRT as it is in travelling to Medelin to try their (fully functional) gondola system. I'm not a non-believer, honest.
  6. @ Peter, you beat me to the point, and I have seen proposals for 1 second headways. I stuck with 2.5 because I could think of a proven system with that ability. Thanks.
  7. Come on. Masdar right now is a few big buildings and that's it. If and when it will be a small city I'd like to see the system working and then I'm interested in numbers. Right now... it's just not what it is meant to be, yet. I'm really looking forward to that day!
  8. Scott, the Masdar PRT system went into operation in November last year. You can view video of it here http://www.prtconsulting.com/prtprojectvideos02.html
  9. @ Steve, Cabinentaxi was developed by the West German government in the 70's, then shelved for political reasons. Ther certifying authority would have been the West German Ministry of Transport, or some such.
  10. Thanks Steven, I just wanted to clarify things a little. PRT has been a field of broken dreams for a long time, but not all of the criticism is deserved. As for Google, their technology is impressive, as are the 'stealth' testing methods they used. The biggest hurdle for them, unfortunately, is regulatory. Who's responsible if a robotic car crashes on a public street? If you go for designated 'driverless roads', isnt' that PRT all over again? IMHO, the only sensible solution is zip lines ;-}
  11. Steve, I am not trying to be dishonest - I am providing you with actual links to real systems. Had you visited our website as i suggested you might have seen this cabintaxi video http://www.prtconsulting.com/prtconceptvideos07.html The Germans had a very robust PRT development program that was abandoned for reason I do not fully undertsand. Yes, the present Masdar system is small - 2 passenger and 3 freight stations. Yes, the plans have been scaled back, however, I believe they still plan to expand it somewhat. Yes, the other system about to come inot public service at Heathrow Airport is also small (3 stations). However, both these systems and the Vectus system which has Swedish DOT safety certification at 3 second headways have similar theoretical capacities to Morgantown - 20 passengers every 15 seconds = 4 passengers every 3 seconds. Let's talk some more about Morgantown. Morgantown has been running for over 35 years and has completed over 140M injury-free passenger miles - try that in abus or a train! Morgantown had a disastrous beginning but has run very well since the initial bugs were ironed out. It still exceeds transit level of service A and the University is considering upgrading and expanding it. While it is group rapid transit (GRT), not PRT, the lines between the two are becoming blurred. Vectus is developing a GRT vehicle that can operate on its PRT guideway. the concept will be to use GRT between selected station pairs with heavy demand and PRT elsewhere. Modern PRT has proven difficult to bring into public service. However, automated cars will be no less difficult. Operating automated vehicles at speed in an open environment while meeting the very stringent safety requirements for automated vehicles is extremely difficult.
  12. One of the ideas of PRT I find intriguing is that it doesn't really have a trunk line/feeder duality. Many existing transit systems have low capacity, short range transit (bus, car) feeding into some high capacity long range system (rail, subway, etc.). Thus there is this view that PRT needs high a capacity to be viable - we want that trunk line. This view is similar to a high way/local road duality. But in reality, a big street grid has a much higher capacity than the one trunk high way, and that's where PRT can be interesting. Because it should work along a PRT grid. And then the capacity of every individual line may not be so important. 10 parallel PRT lines, along every road, can have a high capacity along them - and they may represent the same area that will feed into a transit line. It gets also interesting when you consider that every road may only need a PRT line in one direction on it. If you get on and what to travel in the opposite direction, the system can have you run for on block in the wrong direction, then have you turn twice to go the right direction along a parallel route. The biggest problem I see is cost - both to implement every route KM - especially given the low capacity per route KM - and that you have to commit to a fairly large system of not fully proven technology to benefit from the network synergies.
  13. Steven, I am beginning to get upset by your conitinued accusations that I am trying to deceive people. In the future, please be polite enough to let me first respond to your rebuttle before making accusations. Yes, the reason Morgantown is safe is because the guideway is fully dedicated. It is also because of the fact there are no crossings (only merges and diverges) and because of the computer control system and automated vehicle protection system. These are all PRT characteristics. Your concern about group rapid transit passengers wanting to alight at different stations was addressed in my sentence "the concept will be to use GRT between selected station pairs with heavy demand and PRT elsewhere." So the GRT vehicles will be used between, say a downtown rail station and an airport, with PRT being used elsewhere. I agree with you that asking cities to commit to a huge network is just not going to happen. As we are already seeing, PRT will start out in campus-like environments - universities, airports, hospitals and small communities. it will be used to stretch the reach of other modes like rail and LRT. Over time headways will come down, speeds will go up and networks will be expanded and linked. All I am asking is that you do not knock PRT without giving it due consideration. PRT is not the solution to everything. Nor is it much of a threat to other modes (at least not for a long time). However, it is a mode that, I believe, fills a need and can help improve the viabilty of other modes with a last mile problem, such as high speed rail.
  14. When PRT supporters start talking about massive PRT networks with multiple tracks down every street, and presumably stations at least on every street corner; it starts to look a lot like the automobile-road-system package, but more expensive and less useful.
  15. Just noting that the number of wasted seats on 4-passenger PRT service sounds like a lot - 2-3 wasted seats per trip. Suppose the vehicle gets 100 MPG equivalent (they are electric), then at a loading of 1, the passenger is getting 100 pas-mpg. By comparison, a bus with average loading of 10 that gets 5 mpg is getting 50 pas-mpg. So, simply declaring that we cannot afford to waste seats is a spurious argument. The more important questions are - if we DID implement PRT widely and those seats were wasted, would the total energy use and/or cost go down?
  16. Given that Google's technology (let's call it robocar) is being held up as the shining beacon, let's do some comparative analysis between it and PRT. The primary difference between PRT and Google's technology is the lack of dedicated, grade-separated right of way for the vehicles, so we should accept that as an inherent feature of robocars (otherwise, you get PRT). But there is a reason why dedicated, grade-separated ROWs are a desirable feature. Let's go over some of the considerations: 1. Speed ------- Robocars will not be able to operate at higher speeds than cars presently do on city streets, simply because pedestrians, animals, etc. can at any time step onto the street. On highways (grade separated), the reliance on traditional traction (rubber tires on imperfect asphalt) means that increasing speed will increase the risk of catastrophic accidents in the event of tire blowouts or other mechanical failure. This is especially so if there is mixed traffic with non-robocar vehicles, which is likely. Inclement weather and deterioration of traction will also necessitate slower speeds. 2. Capacity ----------- For the same reasons I mentioned in the 'speed' section, BWS criteria will need to be applied to robocar. This means 3 second headways, which is actually less than typical operating headways on smooth-flowing highways (which is about 2 seconds). Headways for robocars are inherently no better than PRT. If, as you suggest, 5 second headways are required, this will represent a substantial reduction in road capacity. For city streets, since crossings are at grade, each direction of traffic gets priority only half the time on average (and typically less to allow for left turns and signal changes). This means for a system operating with 5 second headways, vehicle capacity is at most 1 vehicle per ten seconds per lane on average, or 360 vehicles per lane per hour per direction. You could theoretically (with a very sophisticated traffic coordination scheme), manage it so that platoons of robocars arrive at an intersection just as their direction gains priority, and thus reduce stop-and-starting and improve average speed. Of course, this requires homogeneity with no user-operated cars, and even then, a single pedestrian stepping into the street at the wrong moment can cause a ripple effect requiring vehicles to miss their appointed windows at intersections. 3. Cost ----------- Robocars will require sophisticated sensors and much beefier computational power than is currently available in cars today. It's not a mere matter of adding software. This will inevitably add to the cost of each vehicle. Fortunately, robocars would make it possible for the rise of robotaxis (which more closely resembles PRT than private robocars), which would be much cheaper than current human-operated taxis. This has the benefit of drastically reducing the number of vehicles required to be owned by the population, saving a substantial capital cost. Unfortunately, there will still be many private robocars, and it seems pretty inevitable that a robocar system will have many more vehicles than a PRT system of similar capacity, representing a larger capital cost. Even more unfortunately, these vehicles need to be stored, which can be quite costly in dense urban areas. Roads aren't free, either. This is a cost that's usually borne indirectly through society, but increased car traffic requires investment in new roads, not to mention the maintenance of existing ones. Many robocar advocates like to claim that the infrastructure is ubiquitous and free, and this just isn't true. The vehicles will also be heavier than PRT vehicles could be. PRT vehicles can get their energy from the grid, whereas robocars will have energy storage (either gasoline or battery). They must also be much more crashworthy than PRT. Thus, they will require more energy to operate. 4. Safety ---------- Robocars are inherently less safe the grade-separated transit. Pedestrians and cyclists will continue to be struck by cars, given at a substantially reduced rate. Robocars will likely provide a large improvement in safety over cars today, but will fall well short of the safety possible in a true PRT system, particularly captive bogey systems. There are probably other factors we could consider. Don't get me wrong--I think robocars have potential. I just don't bellieve they will obviate the need for better transit options, and that really ought to include PRT. PRT can easily outperform traditional transit and robocars in cost, speed, and efficiency in cities. Robocars will be very useful in the outer suburbs or any area with too low density. Heavy transit like subways will continue to be useful in very dense areas of cities. Buses and light rail better watch out, though, as well as a large proportion of car trips.
  17. Steven: 1) Investment. There's nothing inherent in the technology that says it can't be done. Google is where it is (a prototype that is not ready for unsupervised use in the general public), because they have access to large sums of money and highly talented software developers and engineers. Good on 'em. I'm not even convinced that Google intends to commercialize this technology. 2) If you put robocars in a dedicated ROW, you still have all the capacity problems (@ 360 vehicles per lane per direction per hour). You allege that these make PRT infeasible, therefore robocars are infeasible. I reject your headways assumptions in general because they are ridiculous. Human drivers operate at 2 second headways. My point is that you can't have it both ways. If PRT can't pull off better headways, robocars certainly can't. 3) Grade separation isn't a minor modification. Grade separation means different elevations--either elevated or below-grade. That means very large expenditure, and probably higher cost than PRT could do it (only because a robocar PRT is not the most efficient instance of the paradigm). Because the vehicles are fairly heavy, the roadways have to be built to carry substantial weight. Without grade separation (and the substantial expenditure associated with it), then my arguments are still very much relevant. In anticipation of your disagreement on grade separation, please explain in greater detail what you mean by it, and how it would be accomplished, as well as how extensive grade separation will be, and what impact it will have on other forms of transportation in the city, including pedestrians, cyclings, heavy trucks, emergency vehicles, etc.
  18. "the PRT industry can’t even get a 2 mile long, 3 station system to work?" The Heathrow system works by all reports. They have been slow in opening it to public operation. But real live people have been using it. ULTra is in negotiations to install several systems in India. I guess that will be the proof that some will need that the technology works. ULTra is pretty far from the ideal, for PRT. I see it like the Model T. But it will be good to see the technology get some serious attention, and hopefully some more investment.
  19. Alright--we're getting somewhere. Curbs aren't sufficient for grade separation, particularly at intersections. So, it's effectively no grade separation at, especially since an outside vehicle could hop the curb quite easily if there are no bollards. And because we're not using the third dimension, we can't add any substantial additional capacity to existing cities. I think we'll also be able to agree that robocars won't be able to travel at higher speeds, and thus substantial increases in average speed in cities is unlikely. PRT is not in mixed traffic, and thus can travel at higher speeds without endangering other modes of transportation. Perhaps that additional speed is worth something. What, precisely, is the benefit of robocars if we don't get increased capacity, energy efficiency or speed? Is it just safety? Safety is a laudable goal in itself, but it will hardly make robocars a disruptive technology alone. You're heralding robocars even though we only have prototypes (you can't buy it), while dismissing PRT as vapourware when we have several installed commercial systems (you can buy it) carrying real live passengers. Forgive me for thinking that you're more partisan than you claim.
  20. Steve, you seem to be overlooking tha fact that 2getthere is open to the public. While the system only has two passenger stations, it also has three freight stations so it is demonstrating mixed passenger and freaight operations on a five-station system. Also it is disingenuous to claim that PRT technology cannot be made to work. Morgantown can operate in scheduled and on-demand modes which actually makes it more complicated than PRT needs to be. The technolgy is not the problem. The fear of trying something new in an environment (public transportation) that is seldom profitable and that typically involves numerous jurisdictions is the real problem. Developing a new mode of transportation that requires new infrastructure is no mean feat in this environment. The problem with driverless cars is where the liability rests - surely not with the driver. So the company takling the liability is going to have to assume the worst (for liability as well as regulatory reasons). So in snow or freezing rain, they will have to assume the worst possible friction is available. This means very slow speeds and high headways in these conditions - in other words severe congestion. Morgantown PRT, on the other hand heats the guideway and is usually the last system to shut down in adverse weather - and they do so because nobody can get to the system. ULTra has its own methods of dealing with adverse weather (succesfully demonstrated when the last snow storm shut Heathrow Airport down. 2getthere is protected from the weather in Masdar. Vectus has demonstrated full speed operations in snow and ice (the vehicles scrape snow and ice off the rails). Dealing with adverse weather is simply much easier on a small dedicated guideway. This is an important issue most robocar enthusiast overlook.
  21. True enough, Steve. These topics are worth hashing out. I suspect you attracted attention because this post is essentially a hit job on PRT. You can't fault people for quibbling with your assertions. Either you want a debate or you don't.
  22. There ought to be someplace where people can seriously discuss all the crazy ideas in transit technology.
  23. There is a place to discuss crazy, new, and other ideas: http://www.advancedtransit.org/news.aspx ATRA sponsors this active discussion list: http://groups.google.com/group/transport-innovators
  24. @ Steven, This is one thing I truly hope you are wrong about. Though I would love to see driverless cars and expect that they will be inevitable, I really hope PRT comes sooner than later. One thought I have is about a CPT/PRT hybrid. This seems to be extremely logical. You wouldn't have to worry about headways for safety as they would be cable propelled and therefore would all travel at the same speed and react at the same time.
  25. @ Sean, at least theoretically, it would be possible to build a PRT like system using MDG or BDG technology, as they both allow gondolas to be detached from cables, and thus routed to another cable going somewhere else. I would worry about the resulting uneven loading though.
  26. @ Scott, Actually I meant terrestrial which would be much easier to implement than aerial.
  27. @ Steven, A forum would be nice, but if such a thing exists elsewhere, it's not really necessary to re-invent the wheel. I wouldn't mind discussing the craziness of a CPT-PRT system. Sam Wong in the Thought experiment thread mentioned a manually operated variant easily implementable in the developing world.
  28. Would those Station Attendants be the same ones that force us to "Buddy Up" in our automoniles? PRT needs to be considered in an area network context, random origin-dstination pairs. Not just replacements for the several fixed sements mass transit has to use to geet somewhere in the real destination.
  29. Steven, I have been thinking about your arguments and I realized that they apply to cars and roads too. I started doing some calculations and looking up research and it turns out there is no way we can afford such low occupancy in cars. If we assume 1.2 people in a car, and the car occupies 20 feet when stopped and much longer when moving, it would require about HALF OF THE ENTIRE LAND AREA of a city, just for transportation! Even then, the at-grade intersections would be backed up and it could lower average speeds to under 10 mph. That's obviously ridiculous; no city would even consider that - people would stop going to work. And the accident rate would be ridiculous - cars breaking down all over the place. Energy use of such a system would be astronomical. Computer modeling of such a system show that it cannot be done - the needed capacity just isn't there. I think you should start a blog on dispelling the myths of the car system and make sure people realize that IT CANNOT BE DONE before it's too late.