Are London Heathrow’s ULTra Pods the Future of Transit?

I think PRT is an excellent way of solving the “last mile” problem faced by transit planners. It is ridiculous to think PRT could replace higher capacity urban transit (light rail, subway and certain bus lines) in certain areas like the inner city. But to take people from a transit stop to close to their home, or for inner suburban type travel, such as home to town center, schools (think about how much waste comes from diesel school buses in many suburbs and cities), shopping centers, etc, PRT could make sense. Roads themselves can only handle so many cars, and taxis are less than convenient outside of most inner city areas since you need to call for them and wait. People aren’t willing to do that.

My main problem with PRT is the track. The idea of elevated tracks everywhere is not appealing, but I’m not sure other options would work, or be accepted.

I was hiking along the rivers edge in Cardiff earlier this summer when I came across the test track for the ULTRA PRTs pictured. I was quite weird to come across it in such a setting! I watched them for a bit circle around the track and into and out of the stations. Yes – there is a separate siding for the stations, allowing other pods to get by if not stopping there. The cars were quite quiet – just the hum of the wheels on concrete as they passed. The cars I saw got up to maybe 20 mph (I would guess), slowing as they pulled into the station.

For an airport like Heathrow, it would probably work well, as the terminals are scattered across the property in a non-linear way (unlike – for example – Denver International).

For more on Personal Rapid Transit (videos, links, studies): http://www.prtstrategies.com.

Something like the Aerobus technology would make for cheaper elevated track, with the pod suspended from light rail track laid on suspension cable. On a PRT system rather than a LRT like system, the stop could well be at ground level with the stop siding dropping down and then under the through track.

A six person capacity at 35mph with only requested stops would make for a fine local-transport and last-five-miles system, feeding into a higher capacity rail backbone.

It may be that part of the over-promising of PRT advocates comes from trying to sell it as a one-size-fits-all transport solution, even though we are still laboring under the burdens imposed by the last big assumption that a one-size-fits-all system actually exists – as the car-based system proved, yet again, that one-size-fits-all fits many poorly and never actually does fit ALL.

Interesting article. Just wanted to note that it’s West Virginia University, not the University of West Virginia.

Michael, re: PRT connecting to transit: this resembles the way some neighborhoods are served in Hong Kong and Singapore. In those cities, the subway network is not dense outside the central regions, and skips many neighborhoods and housing projects; instead, those neighborhoods are served by airport-style automated people movers, which are called light rail even though they’re technologically and operationally different from what is called LRT in the West. Those people movers aren’t PRT, but it’s conceivable something like PRT could be used for other areas not yet connected by people movers.

The main problem is that PRT doesn’t really work with high-traffic destinations, such as train stations, shopping centers, office buildings, and schools. Access and egress times are fairly high – I’m guessing on the order of 10-15 seconds – which means that the capacity of a single platform is about 300 people per hour. Even vehicle sharing couldn’t help much, because if you put 2 people in a vehicle side by side, then access and egress time would double unless you had doors open on both sides of the vehicle, which would increase station footprint.

Mike: I’m not sure why you think that American light rail cannot achieve full capacity because transit agencies are incompetent, but PRT would work well. Incompetence goes both ways.

@Alon:

Regarding station capacity: a large capacity station will have more than one berth for PRT loading/unloading. In that case, the 15s load time per vehicle would occur simultaneously in N separate vehicles.

Further, ULTra’s station berths are arranged in a “sawtooth” form so that all vehicles have independent access to the exit guideway. This ensures that a single “slow loader” doesn’t back up the whole station, as would be the case with a simple linear arrangement.

So consider a station with 8 parallel berths and 20 second turnaround per berth. That equates to 8*3*60 = 1440 vehicles per hour. This platform would be about 4m*8 = 32m long. And the highly parallel nature of the ULTra platform means it could probably expanded to more berths if needed.

And I never said US transit agencies are incompetent, far from it. The US suffers because of cultural dominance of the automobile, which is the result of various social, political and economic factors over the past 100 years. Transit agencies were the victims of this centuries-long paradigm shift, not the cause.

Thanks, Mike C, for the additional numbers.

That would mean, that a “siding” (access to a single house in the “suburban” scenario) would be about 100 meters long.

Hmm… that pretty much corresponds to the length of the ramps to and from a house for ground level access.

Hmm… that would give a minimum distance between individual house stops of something around 70 m (with some smart intertwining of the access ramps. Now, I wonder whether a suburban area with a distance of 70 m or more between houses would be willing to accept such a wall of access ramps. I don’t know…

One stupid question concerning the “sawtooth” or “parallel” berth idea: how are the passengers supposed to get to and from the berths? Assuming that the berths are not stub ends (which they aren’t, because of the term “exit gateway”), the “platforms” are insular platforms, and we have established before that the guideway can not be crossed by pedestrians because of the height of the guide bars.

Ahem…

Actually, it would be possible to have one long curbside, IF the vehicles are capable to operate without guide bars for a short distance, and with one single guide bar for a bit more. Such a stop would look like a bus bay.

Another aspect, I wonder how it will be solved in Heathrow, is the handling of empty vehicles.

For one, if it was a bidirectional track, that increases your calculated capacity by a factor of 2.

Could you explain this? I don’t see how bidirectional tracks double capacity. Do you mean having two tracks and two platforms? If so then yes, the capacity is 300 people per direction; the problem is that demand is often asymmetric. You can only really capacity per direction if you’re at a terminus, which may or may not be the case; this is equivalent to putting two platforms serving a direction.

Offices don’t have 300-600 people leaving at one time, but they do have 300-600 people leaving within an hour, which can be met even by automobiles on local streets. What cars can’t meet, PRT can’t, either: large schools need to accept or discharge 3,000 students within 20-30 minutes. City buses can meet this with multiple stations on multiple streets, and rail has the capacity for this, but PRT would have queuing nightmares.

Not all transit stations have enough footprint for the 10-20 platform tracks you’d need to meet peak capacity. About the only way I can see of making this work is dispersing demand, which means having PRT run alongside mass transit for 2-3 stops. For major destinations you’d have multiple stations for different sections – though even then PRT would have problems with track capacity, and multiple stations make more sense for large-footprint factories than for schools.

@Alon:

Yes, PRT designers generally envision a larger number of very small stations scattered about. One of the more promising porposals is to actually incorporate stations into buildings themselves, and then it truly does becomes something like a “horizontal elevator”. This would reduce costs and enhance service, and I would think that property owners would be eager to provide such a convenience to enhance the value of their property.

My point about long lines of parallel station berths was not that it would be the norm, but rather, that it would be possible if needed.

For the “600 people leaving within an hour” problem, consider a 5 berth station near the building. At 15-seconds turnaround, that’s 4*5*60 = 1200 veh/hr. A 5-berth station is about 75 feet long, which sounds like a lot until you consider that 600 people are no longer parking their cars near the building. 600 parking spaces is 13,200 linear feet of parking (at 22 ft/space). If even a dozen of those parking spaces can be reclaimed for PRT stations, that’s room for plenty of PRT berths to service the required load.

One more point about PRT guideway capacity: as the network expands, overall capacity expands with it. PRT’s ability to use the entire 2-dimensional network means that local surges in demand can be spread out over a larger area. For example, consider a concert venue with a 5000-person capacity. That venue could be served by a number of PRT stations around the perimeter of the venue, each of which can take riders out via a different route. With a dense PRT network, there are many possible routes from point A to point B, so those 5000 concert patrons might be cleared in significantly less than an hour even though no individual guideway segment carries more than 1500 vehicles/hr, because the routing system automatically disperses vehicles to alternate routes when the primary routes are full.

For example, consider a concert venue with a 5000-person capacity. That venue could be served by a number of PRT stations around the perimeter of the venue, each of which can take riders out via a different route.

This doesn’t work, not when construction cost is $14-30 million for a two-track route mile; in that case, you could have a reasonably dense network, but it would be rare for a destination to be served by more than one station or maybe two. City buses can sometimes handle loads at concerts because the street is already there, and isn’t very expensive to build or maintain, so you could have multiple lines. But as soon as you hit light rail costs, you have to have the same economics as light rail.

@Alon,

The guideways and stations would not necessarily serve only that venue. What I was suggesting is that the larger the network gets (in both area and density), the greater the capacity to overcome local surges. So in my example I said stations would be located around the perimeter of the venue, but they would also serve adjacent buildings, and there would have guideways going off in different directions to cover different areas.

Think of a 2-dimensional grid with the concert venue at some point within that grid. At a minimum, there would likely be 2-4 guideway paths away from the venue as part of the overall network layout. These can be used in parallel to multiply traffic out of the venue. Once out of the congested area immediately surrounding the venue, vehicles can be routed efficiently down the least congested path towards their destination.

Since a concert venue is a known aggregator of potential passengers, the overall layout can be optimized to have several guideway loops converge at or near the venue, even as the primary purpose of those loops is to serve surrounding areas.

As for costs, I frankly don’t understand why PRT turns many transit activists into fiscal conservatives. Yes, the system is expensive, but so is rail, and transit activists don’t generally object to well-conceived rail projects. PRT is expensive to build but it would provide a valuable service to the community, like any transit. And if it is heavily used by a large percentage of the public, capital costs for expansion will be less controversial. I can’t remember the last time I heard a public outcry over road expansion/maintenance costs in my car-dominated city, because people view roads as a vital service. I think the same will be true of a well-conceived PRT system if a significant percentage of the public embraces it — and I think they will once they see the high level of service it provides.

“Buses (with the exception of those that use bus-only lanes) cannot provide efficient transit.”

Please to explain your logix, Mr Geoff.

Mike C, #78
This means that there is only a one in 10 million chance that the follower vehicle will NOT have its brakes engaged within 100 milliseconds of the fault in the leader.

But the chances of programmer error in some part of such complex software is one in one. I’ll stick with physical redundancy, fail safe operations and safety, thank you.

Also, EXACTLY what do you mean by 1 in 10 million chance of failure. Is this per computer calculation or something else? You’re a programmer; please explain in plain English so the non-programmers here can understand what you’re saying.

Yes, Alon, the track record for the Shuttle program is 1.55% failure so far, based on two disasters and 129 missions through July 2009. I think they’re retiring the program because NASA accepts the fact that a 3rd Shuttle disaster is inevitable.

I also know that the chance of Word crashing, even on a Mac (which has an extremely stable OS) is one in one from “out of memory” errors (I think this is “memory leak” in programmer-eese) if I leave Word open for more than 24 hours. The current Mac version of Word is much better than before, but it still sucks relative to virtually all other Mac software.

@Michael Setty:

OK, so now you’re comparing PRT to attaching a reusable vehicle to several tons of liquid explosives, launching it into orbit, then returning it safely home? And this is supposed to predict the reliability of operating golf carts at 25mph on an exclusive guideway?

Really Michael?

And I wasn’t directly comparing automated golf carts with Shuttles. You misrepresent what I said, again. Ho-hum.

On the other hand, a PRT system as a whole probably would be comparable in complexity to a Shuttle. Of course, I doubt you or any other PRTista has any clue where that point of complexity is. Clearly, the Heathrow PRT is at the low end of the evolving PRT complexity scale, and they’re still have major teething problems that recently led to a nearly year long delay.

Also, are you saying YOU have a thorough understanding of reliability engineering? Sounds like it. Am I supposed to take your vast knowledge of aircraft reliability engineering on faith? Oh, wait. I’m also to have “faith” in the very low infallibility of PRT technology.

Alon, you beat me to the last comment I left out; e.g., that the latest, most complex jumbo jets have also had the benefit of 106 years of airplane flight testing and myriad other efforts since the Wright Brothers. In contrast, the history of PRT and AGT is rather sporadic in comparison.

@Michael Setty wrote “On the other hand, a PRT system as a whole probably would be comparable in complexity to a Shuttle.”

Michael, the most complex PRT system is a tiny fraction of the complexity of the space shuttle. PRT is not rocket science — literally!

ULTra PRT is basically a golf cart on a guideway with a magnetic tracking system and obstacle detection. The central control is little more than simple linear trajectory calculations (which have been automated since the days of ENIAC) coupled with some routing and scheduling. From a pure functional perspective, that’s it. The role of the PRT engineer is to make this handful of simple functionality reliable, and judging by the regulatory acceptance of several designs, they’ve succeeded.

The fact that you believe a PRT network would be anywhere near as complex as the space shuttle reveals your fundamental lack of knowledge on the concept.

RE: PRT vehicle reliability

Any PRTistas here know what the projections are for miles or km between roadcalls being predicted by any of the PRT systems under development?

If a bus system keeps this measure above 10,000 miles between roadcalls, they’re doing reasonably well. Buses are relatively simple, of course, but they still do break down, and there is a long empirical record of why and how.

Bruce, you’ve done more to marginalize your position than even Mike C. That’s quite an accomplishment.

PRT proponents have struggled to build a viable working model. The question is not the computing capabilities required to fulfill PRT designers’ wishes.

There are a host of real-world operational issues involved that are beyond the scope of solving by silicon chips.

Setty brings up a very good point about maintenance. What will be the forecast of road calls? More importantly, who will be responsible for them? Also, what is the schedule for routine/preventive and major overhaul fixes? Where are the parts’ sources?

There’s also the issue of operational costs. What is the funding plan to keep the pods and guideways running? Ignore the fact that there is theoretical savings from not needing drivers. Look at the costs that are there. What will pay for PRT’s power source? What about the cleaning and maintenance workers? Better yet, what is the revenue source? Fares? Even the most efficiently run systems operate at massive losses. PRT’s “fuel” cannot be taxed. Or will it be like a sidewalk, which has a farebox recovery of 0%?

What about system redundancies in the event of service failure? What’s the backup power? What are the alternatives available should all or significant parts of the system be out of commission? Bruce, we’d look like asses if PRT make mass transit obsolete — until the point comes when we need obsolete technology to rescue riders from a major service disruption.

A more important question concerning vehicle reliability–what happens when one does break down? If there’s a “grid” topology, it theoretically can be routed around, but a linear route would be shut down by such a happenstance, would it not? A bus breaking down doesn’t take out the entire bus system; a train breaking down on a properly dual-tracked line can be routed around (in worst case, a broken down train can be bridged with busses if routing around it is not possible).

@Wad:

“Bruce, you’ve done more to marginalize your position than even Mike C.”

I’m not interested in whether my position is marginal, or orthodox or not. Never have been. However, I will say that my attitude to trains and buses is very widespread — to the extent that those who take the opposite view are the unusual ones.

“Bruce, we’d look like asses if PRT make mass transit obsolete — until the point comes when we need obsolete technology to rescue riders from a major service disruption.”

That’s a hilarious statement, in light of my experience of public transport. I live in the UK. Trains here are regularly disrupted, sometimes by technical faults, sometimes by “planned maintenance”, and sometimes be industrial disputes. Then buses are used to provide a replacement service. Buses, too, are prone to being disrupted by industrial disputes. Given the proven unreliability of these services, it is a joke to say that they will be needed to step in as back-up for a system whose unreliability is not yet proven.

@BLambert:

“I don’t like transit, so I want a car to take me where I want to go.”

No, I don’t want a car to take me everywhere. I find driving mostly a boring activity. A year after I bought my first car, a colleague asked me why I always walked to work (a trip which took half an hour). Cycling I enjoy. I’d like safe cycle lanes, and I think covered cycle-ways would do more for green transport than suburban trains or buses ever could.

That said, if PRT works as well as it is theorized that it could, I would much prefer to ride the PRT than a train or bus. My reasons for disliking buses and not much liking trains are well-founded and widely shared. PRT is transit – it moves people around, and it is shared by the riding public. If I like it, then I don’t dislike transit as such, though perhaps I dislike mass transit. And why should I not? Noisy, crowded vehicles in which one is often forced to stand squashed like a sardine, shockingly slow travel, long wait times, service inconveniently discontinued at night and sparse on weekends – what not to dislike about mass transit? They are at best to be stoically tolerated.

I have no problem with fast, smooth, clean, comfortable, always-seated trains with reasonably low interior noise, such as the TGV and the Shinkansen, and I sort of understand people who want to ride unusual and very old historical trains, but to like these forms of travel in general – to find anything at all to love in commuter rail, metro rail, city buses – what kind of weirdo does that?

If a better system can replace them, well, hooray. Why should I care if some people think PRT is not transit? It promises quickness, convenience and comfort, and all at a considerably lower cost than trains. That’s what matters to me.

@EngineerScotty:

“A bus breaking down doesn’t take out the entire bus system; a train breaking down on a properly dual-tracked line can be routed around (in worst case, a broken down train can be bridged with busses if routing around it is not possible).”

In my experience, a broken-down train generally shuts down an entire line for anything from a few minutes to a couple of hours. A broken-down bus is not such a serious problem, but roads can become gridlocked for all sorts of reasons, and this does affect buses.

Obviously, PRT designers have thought about the problem of broken-down vehicles, and have proposed measures to deal with that problem. And obviously, since you’re an engineer, Scotty, and you’re interested enough in this topic to be posting here, you’ve read that one of their suggestions is to design things so that any working vehicle can be employed to push a broken-down vehicle to the nearest station. That seems like a reasonable solution which would prevent a widespread service disruption from occurring in most breakdown situations.

I’m not interested in whether my position is marginal, or orthodox or not. Never have been. However, I will say that my attitude to trains and buses is very widespread — to the extent that those who take the opposite view are the unusual ones.

Bruce, your position isn’t widespread, except maybe among suburban NIMBYs and oil industry shills. Across the Sunbelt, cities are voting to tax themselves to improve bus and train service; in LA County, the margin was 2 to 1. It’s not very useful when you complain about facts that are “easily checked and shown to be false” without actually showing why they’re false.

And I’m glad you’re not interested in what effect you’re self-marginalizing. But that raises the question of why you’re writing comments in the first place. Is it just to tell everyone how great you are that you don’t like waiting on subway platforms? Or is it to tell everyone that you’re smart enough to know that the original spelling of utopia was eutopia?

Re reliabiliity: Here is ATS’s discussion of reliability and how to handle vehicle breakdowns.

Short answer: the vehicles are engineered to very high reliability, with automated vehicle health monitoring systems to trigger maintenance on issues before they become severe enough to cause an outage. In the rare event of a breakdown on the guideway, they have procedures in place to handle it.

Why is this so difficult for some to accept? Do people here really believe that ULTra could have gotten regulatory approval without demonstrating vehicle reliability and contingency plans? Again, ULTra was designed to aircraft-level reliability standards; how many people here worry about planes suddenly dropping out of the sky?

I’ve been involved in online PRT debates for years now, and it’s always the same: people who’ve never even read a single page on PRT suddenly think they’ve found the “killer fault” that invalidates the entire concept, despite the fact that many very intelligent people have been working on this for decades and have analyzed to death every conceivable scenario.

And I want to stress again: regulatory agencies have approved PRT for carrying passengers. So those here who think it’s up to the “PRTistas” to prove safety and reliability have it backwards. Regulatory approval is a daunting and comprehensive task for a public transit system, and it covers all the issues raised here and dozens more. So for those who distrust the regulators findings, the burden of proof is on you to find where they went wrong.

@Max, it’s an emergency walkway which will be used exceedingly rarely (if ever). Do you also question why the inflatable slides used in airplane emergency exits don’t have side railings?

I’m reminded of this exchange from Raiders of the Lost Ark:

-“We have top men working on it right now.”
-“Who?”
-“Top… men.”

PRT looks like the best bet to attract new private investment in public transit to help address congestion and mobility.

Do you have any evidence? Or are you just assuming that every promise of PRT is true and every promise of mass transit is false?

The private sector has some participation in bus. This participation is mainly limited to shuttle buses to and from rapid transit stations. It also includes assorted other shuttle buses. The remaining private participation is as a subsidized operator. Veolia et al typically run the service, turn over fares to the transit agency, and receive their revenue based on meeting their performance criteria. This contractor relationship does not constitute “investment” per se. The last example is of toll roads and HOT lanes where the concessionee runs a bus service or provides a fixed guideway as part of their contract to build/operate the toll facility. Again, a pretty limited investment.

The private sector has much more limited participation in rail. Mostly, this participation comes at the station level – redevelopment partnerships where the developer improve the station or more often the parking supply for the station in exchange for development rights. Value capture financing is a form of soft private participation. Both of these elements are pretty weak and do not go very far towards footing the bill for grade-separated rail.

On the other hand, people movers have seen substantial private sector influence. PRT in my view is just the next logical step beyond a people mover. Very similar technology with a further degree of complexity. As far as I can tell, it allows for a plug and play approach where a new investor will pay for an extension loop on the network as a real estate enhancement. In some place like Silicon Valley, this market is potentially quite large. The incremental investment for each loop is more modest, too. And a unsubsidized capital investment that is then turned over to a transit authority would thankfully skip a lot of state and federal red tape – every thing from Davis Bacon requirements to prolonged environmental review. The cost basis is looking pretty favorable. In the mean time, it would provide a much better experience to the customer than traditional transit, which most people just don’t want to pay for (yes, even those who do).

Transit is like tofu. Sure some people eat it because they like it, but in the aggregate, we all prefer pork, or lamb, or beef, or even frickin’ chicken. You can keep telling me it’s terrible for my heart, the planet, etc… But it tastes better, and I’d rather engineer a solution to keep the habit than to eat your tofu. Because tofu just isn’t that good.

The private sector has much more limited participation in rail.

Veolia? JR? Japanese private railroads? MTR? MRT? Detroit light rail? Tel Aviv subway? Channel Tunnel? Desert Xpress? Each of those has private sector investment higher than the Winona or Morgantown projects. That there’s more private investment in PTR than public doesn’t mean much – the overall numbers are low; it’s just that the public sector is even more loathe to invest in PRT than the private sector.

And a unsubsidized capital investment that is then turned over to a transit authority would thankfully skip a lot of state and federal red tape

This is not true. Desert Xpress, a privately funded scheme, has to contend with the same environmental review process as public schemes like California HSR.

@BLambert:

“I said nothing about driving a car, I said you want a car to *take you* where you want to go. You don’t like mass transit because, horror of horrors, it’s loud and it’s got people in it?”

Er, no. I enumerated my reasons for not liking mass transit, and the mere presence of other people was not one of those reasons. I highlighted the slow trip times, the uncomfortable, very noisy ride (almost deafening, in the case of most metro train services), the crush-capacity crowding which makes New Year’s eve on Trafalgar Square look sparse, and the inconvenience of waiting. So, you’re attacking a straw man of your own manufacture.

“That describes cities, too, so maybe you’d be more comfortable not living in one.”

As a matter of fact, surveys have found that people who live in big cities are on average quite miserable compared to people who live in small towns, leafy suburbs and rural districts. I don’t doubt that commuting by mass transit is a contributory factor.

@Alon Levy:

“Bruce, your position isn’t widespread”

My view is very widespread. The overwhelming majority of people in developed countries use public transport rarely or not at all, and indeed spend thousands of pounds to avoid doing so (i.e., buying, insuring, parking, fuelling and maintaining a car). Even in Japan, where the trains are famously efficient and punctual, the majority of journeys are not by public transport.

“It’s not very useful when you complain about facts that are “easily checked and shown to be false” without actually showing why they’re false.”

Others have already done so. I do not need to duplicate their work.

“But that raises the question of why you’re writing comments in the first place.”

(a) Because I have an opinion that is on topic, and (b) because I noticed some bad arguments, and bad arguments deserve criticism.

“original spelling of utopia was eutopia?”

“Eutopia” is not simply an old spelling of “Utopia”. The meaning is different. “Eutopia” means “good place”, whereas “Utopia” means “nowhere place”.

@John W

“How do you evacuate from one of these things if you’re stranded in the air halfway between two pylons? With no guardrail, the side door seems to be out of the question.”

This is another skeptical question that arises from not bothering to do the most elementary sort of research, such as reading the literature provided by the credible vendors. To take the case of ULTra, the front and back panels can be opened in an emergency, and the guideway is walkable (and has a guard rail). If you don’t check your facts, your attempts to argue against PRT, far from carrying weight, will merely discredit you.

“(eg fallen tree branch wedged into the track). Big cranes to lift out the cars?”

You are indulging in special pleading. If a tree fell across a road or a railway line, service would be severely disrupted until cranes were deployed. Ditto, if it falls across a PRT guideway. You’re arguing as if somehow you think that your argument applies against PRT, but magically doesn’t apply to other modes.

@Max Wyss

“… yeah, 6 meters above ground with a 25 cm high guard rail…”

Yet another false assertion that is contrary to the obvious and easily checked facts. The PRT guideway has a guard rail that is about 4 or 5 ft high, as documents explain and numerous pictures show.

@Charlie Dube:

“The more successful transit is, the more unbearable it becomes. It only takes one crazy or inconsiderate person on the vehicle to make everyone miserable.”

Totally agree. People use public transport when they are forced by circumstance. When the level of congestion on the roads and in parking gets unbearable, public transport ridership rises until the road is bearable, and then it stops rising — the lack of comfort and speed, as well as the inconvenience, crowding and the risk of uninvited encounters with nutters all being matters taken into consideration. Hence, the majority of cities in the developed world are unable to boast that most commuter trips are made on public transport, let alone most trips.

@Aron Levy:

“Or are you just assuming that every promise of PRT is true and every promise of mass transit is false?”

It does rather seem to be the case that most of the promises made by mass transit are false. When projects are announced, the promised cost of implementation is usually half or less the final cost. Promised ridership and farebox recovery, by contrast, are usually much higher than is realized in the end. If PRT is merely no worse than mass transit at delivering on its promises, it will be a better system than mass transit.

@Michael D:

“Again, costs are completely ignored. If your density cannot support even reasonable bus service, then how the hell can you justify building a PRT network?”

If you look at ULTra’s website, you will find details of sponsored and independent studies that compare their PRT against buses, and their PRT comes out better on both costs and services. It might not apply in all situations, but it looks as if PRT can out-compete buses in a wide range of applications.

“privacy has its downsides. If we’re talking urban transit, those pods will get nasty. Smoking, litter, etc.”

London buses – no privacy, but when I have ridden them, I have observed filthy interiors and appallingly inconsiderate passenger behaviour. So, if it’s a choice between a filthy private vehicle and a filthy crowded one, I guess the filthy private one might just win by a nose. It’s like being miserable and rich – better than being miserable and poor.

@Aron Levy:

“Veolia? JR? Japanese private railroads? MTR? MRT? Detroit light rail? Tel Aviv subway? Channel Tunnel? Desert Xpress? Each of those has private sector investment higher than the Winona or Morgantown projects.”

You are listing exceptions. Only a tiny fraction of rail services around the world are privately owned and operated, and even most of those are dependent on infrastructure that was built at taxpayer expense. Others that are ostensibly privately run in fact receive substantial state operating subsidy, on top of the state funding of infrastructure.

My view is very widespread. The overwhelming majority of people in developed countries use public transport rarely or not at all, and indeed spend thousands of pounds to avoid doing so (i.e., buying, insuring, parking, fuelling and maintaining a car). Even in Japan, where the trains are famously efficient and punctual, the majority of journeys are not by public transport.

PRT doesn’t even try to handle rural journeys, which is where cars predominate. For urban journeys, transit can and does have majority modal share – for example, rail alone has 60% of the commute modal share in Greater Tokyo. In Singapore, where car ownership is at Manhattan levels, and Hong Kong, where it’s half as high, the rail + bus modal shares are even higher.

Of course, it helps that none of the above cities engages in US-style road socialism. But if $1 million/lane-mile roads need subsidies, what hope does a $7-15 million/track-mile PRT system have? To run profitably, US roads need to double the user fees or more; PRT would need to cost at least 14-30 times more than US roads do right now.

Only a tiny fraction of rail services around the world are privately owned and operated

Except in Japan, which has 40% of the world’s rail ridership, and in most of the rest of developed East Asia…

And the alternative to public transportation, highways, is never privately operated. Local public transportation is private or profitable in many cities; local streets never are, even in countries where mainline freeways are profitable through tolls.

When projects are announced, the promised cost of implementation is usually half or less the final cost. Promised ridership and farebox recovery, by contrast, are usually much higher than is realized in the end.

Promised ridership was higher than realized in the 1980s, when there was no empirical experience in the US – i.e. the same situation as PRT. From the late 1980s onward, US light rail ridership projections were usually correct. For every underperformer like San Jose, there’s a Phoenix, with ridership substantially higher than projected.

If you look at ULTra’s website, you will find details of sponsored and independent studies.

Yes, and I’m sure none of them has a conflict of interest and all were published in top quality journals and are highly regarded by the experts… I’m sure they’re nothing like the sponsored and independent studies touted by the Tobacco Institute.

@Allen,

There is a mountain of difference between riding up Pikes Peak and riding in city streets, pun intended. None of the existing driver assist technologies operates without human oversight, and I don’t see that changing for at least a decade, probably more.

The problem is, when dealing with uncontrolled settings, it is exceedingly difficult to design systems that can handle unpredictable situations nearly as well as a human can. Sure, you can get part of the way there, i.e. alarms that signal a driver when drifting out of a lane or when a collision is pending, but that final step of eliminating the driver entirely is a huge challenge.

My guess is that we will start to see very low speed automated vehicles, i.e. 20mph or less, and mainly in controlled settings. The automated parallel-park is the first implementation of that. But higher speeds and mixing with traffic is a long way off.

The basic problem, in my view, is that automated control algorithms will not be able to mimic the inherent judgement that a human has, not for decades, and therefore automated vehicles will have to operate in a paranoid mode, stopping for obstacles that humans would easily dismiss as inconsequential.

But that’s just my impression, I could be wrong. I have worked on artificial intelligence algorithms in software for more than 10 years (in my case, language algorithms) and I’ve seen firsthand the enormous challenges that must be overcome to build software algorithms which can interpret data as well as the human brain. I’ve seen how easy it is to go from 0% to 60% accuracy in such algorithms, and then how much more difficult it gets after that. I personally think the challenge is exponential — the effort needed to go from 50% to 55% is probably 1/10th that required to go from 55% to 60%, which is 1/10th the effort to go from 60% to 65%, and so on.

So I think we’re decades away from fully automated vehicles on city streets. But as I said, I could be wrong.

Alon:
Surely there’s an institute selling automated cars that will be happy to provide studies for you proving that automated cars are completely safe.

There’s no “think tank” selling automated cars I know of, but there is this site touting “robo-cars” as being just around the corner: http://www.templetons.com/brad/robocars/. Arguing with the fellow behind this site is even more pointless than with PRT advocates.

It’s also very interesting how Mike C understands that the complexity and effort to obtain more competent AI goes up exponentially, as he succinctly outlines, but fails to see that the complexity, potential areas where the system can fail, and the overall potential for system failures ALSO go up exponentially as still unproven PRT networks increase in size, the number of vehicles, and the number of stations.

In a very simple PRT system like Heathrow, there are only two possible origin-destination pairs, “Carpark A” and Carpark B” to/from the terminal station. If this very elementary network operated in a non-airport context, there are only three origin-destination pairs.

But if you go to the next step, such as my theoretical 25 station PRT network for in Winona, MN has a total of over 300 station origin-destination pairs, plus dozens of complex connections between various routes and loops, as well as 200 or more PRT vehicles required, if my traffic estimates are reasonable at 15,000-18,000 trips per day, vs., 1,500 or fewer daily at Heathrow.

Keep in mind that the Heathrow system has been delayed another year, due to various kinds of teething problems–and it is a very simple system compared to the sorts of networks that PRT advocates want developed next.

At Heathrow, the human clientele will certainly be very well-behaved, consisting mostly of well-heeled business air travelers who are unlikely to leave graffiti or otherwise abuse the equipment.

But in a general public PRT network, system operations will be subject to the wild card of the “human factor” including my favorite, the “Little Old Lady in the PRT Station” problem, as well as unruly teenagers, gang members who excel at leaving their marks in unsupervised vehicles (Mike C, have you ever seen the back of an urban transit bus in a large city??), the occasional pervert and drunk, as well as the various other things that currently bedevil public transit, e.g., THE PASSENGERS!

There are many non-technological factors that can upset the automated PRT applecart, which is one of the reasons why it will often be quite impractical to run a PRT system “at the edge” of its maximum throughput without some significant margin of error.

This is a central reason why I’m highly skeptical of PRT. Yes, minimum 2.5 second headways at times may be possible on some systems, such as demonstrated with Cabintaxi 30 years ago, but not reliably at close to the limit for hours on end, when the technological complexities and ongoing maintenance challenges of a highly complex technological system are considered, let alone the confounding myriad of human factors, such as the non-tech saavy little old ladies who “hog” a ticketing/dispatching interface for 4-5 minutes before boarding their pod, assuming they can easily find the correct pod particularly during busy times.

Even when it’s private money paying for a PRT system, big-box transit advocates go blathering on and on…

How many of you have written PRT simulation software, so as to back up your claims that the guideways will get congested and there will be nowhere to park empty vehicles?

Throwing more money at bus and train systems won’t make them successful. The vast majority still choose private automobiles, regardless of traffic conditions.

However, something in-between the automobile and the bus/train, something like PRT, might make transportation–all modes–more efficient… if we finally give it a fair chance to prove itself.

So please, read up:
– Evolution of Personal Rapid Transit

– The Future of High-Capacity Personal Rapid Transit

@Michael Setty:

What “factors”? Larger networks? The algorithms scale linearly; if you have 10000 vehicles instead of 10, it’s simply 10000 CPU cycles instead of 10. But even that linear increase doesn’t matter, because modern computers can already handle PRT algorithms for millions of vehicles without difficulty.

The old granny? She takes longer to board and other passengers just go to another waiting vehicle. Gee, that was difficult. See, the problem is that you’re used to the little old lady boarding a bus, where she would delay the other passengers. Station throughput would be marginally affected, but line throughput would not be affected at all.

Graffiti and Vandalism? Transit authorities have been dealing with these problems for years. The difference with PRT is that passengers actually rent individual vehicles, so vandals can be more easily identified, at least in theory. This should be an effective deterrent, along with CC monitoring; but if not, then it’s no worse than rail transit has had to deal with. And it has nothing to do with throughput.

Breakdowns? We’ve already dealt with that: aircraft-level reliability, self-monitoring systems and preventative maintenance should make service interrupting breakdowns very rare (remember the analogy to planes falling from the sky?), and even when they do occur, there are contingencies as with ANY transit system that fails

What else Michael? You’ve given us little old ladies and vandals and vague predictions of “unpredictable human factors”… and clearly they will not be any more problem for PRT than for other transit systems, and often less of a problem for PRT.

So do you have anything else? I sincerely hope not, because with every post here you show us more of what you don’t know, and honestly I’m starting to feel sorry for you.

[[“PRT doesn’t even try to handle rural journeys, which is where cars predominate.”

So? The majority of trips are taken in towns and cities. Even if PRT never winds up offering rural services, ]]

I meant to say, the majority of trips are taken in towns and cities. Even if PRT never winds up offering rural services, this will make little difference to PRT’s ability to compete with buses and trains, which, incidentally, are incapable of providing good rural services economically.

@Alon:

EDICT report summary, hosted at the EU research site.

The Daventry PRT study, a vendor-neutral study conducted by transit consultants SKM and hosted on a local city government site. I believe the study was government sponsored.

Fundamentals of Personal Rapid Transit – a book published in the 1970s documenting the research done by the Aerospace Corporation, a US government contractor which studied PRT for several years on government funding. Because their work was government funded, Aerospace Corporation owned non of the intellectual property associated with that effort.

Furthermore, Masdar City is deploying PRT on the recommendation of their independent consultants. PRT was selected as the technology before any individual PRT vendor was selected.

The notion that PRT technology and feasibility is tainted by financial influence is silly

Re: Morgantown: no, actually Morgantown runs at 15 seconds headway. Cabintaxi, however, ran at 2.5s and ran fleet endurance testing for over a year, uninterrupted. Do you realize the significance of that? Michael Setty has expressed concerns about PRT endurance for a 3 hour rush period, ignoring the fact that Cabintaxi ran vehicles for thousands of hours nonstop, without fault.

So Morgantown disproves the notion that an automated small vehicle network cannot be safe in a real world replete with Setty’s “grannies and vandals”. Cabintaxi’s trials disprove the notion that a system cannot run at 2.5s headway safely and reliably over a long period of time. Regulatory approval of modern systems provides further validation of the short-headway PRT case.

What more do you need? Independent studies? See above. Real-world validation of modern designs? That’s coming this year (Heathrow and Masdar).

What else do you need to finally stop saying PRT is unsafe/unreliable/incapable of low-headway operation?

Okay, 2.5 seconds… still very low-capacity, 1,500 per direction per hour. Are you going to take back your statement that PRT can be run safely at a capacity of 5,000 vehicles per direction per hour?

I stated a fact, namely that most travel…

…as opposed to urban travel.

You keep dodging.

In Singapore, the total mode share for private vehicles, taxis, and cars + MRT is 32%, according to your link. The balance of the 32% + 55% is made by “others” and “no transport required” (walking, working from home).

As a bit of a newcomer to this debate, I’m a bit puzzled at one thing.

There seems to be some agreement above that mass transit and a successful urban deployment of PRT (should one come to fruition) might serve different niches–traditional mass transit thrives in high-density areas, whereas is better suited towards enabling car-free living in reduced density areas.

Yet many of the comments from PRT supporters above–continue to frame PRT as a replacement for mass transit (or at minimum, as more worthy than traditional mass transit for the limited funding that’s available); and some continue to promote the myth that rail transit is only suitable for cities like New York or Hong Kong, with five or six figure density figures (persons/square mile).

Why is this?

Wouldn’t there be more interest in advocating PRT to developers?

I think that many transit advocates would be less hostile to PRT if PRT advocates weren’t trying to steal their lunch money, don’t you think–especially given that the technology seems best suited for a different role than the bus or the train? Instead, much of PRT advocacy seems to be geared towards diverting the pittance of transportation dollars that goes to rail–rather than getting a chunk of the much larger sum that is allocated towards roadbuilding.

Speaking of money, it also seems to be the case that most of the PRT advocates here have a financial interest in the success of PRT–quite a few of you are working for PRT vendors, it appears. I can’t think of anyone who posts here who works for Siemens or Bombadier or Gillig–the pro rail and pro-bus folks here are mostly transit users and advocates who care about the quality of the service provided, NOT about winning orders for their employer or their stock portfolio.

@EngineerScotty:

I have not advocated for PRT in place of existing rail, and I absolutely do NOT have any financial affiliation whatsoever with any PRT company.

Adding to Peter Muller’s message 148.

We will see how the current projects evolve.

There are issues which will have to be addressed. Among them:

• Costs
As it presents itself, the costs are rather high, for what a system provides. At the moment, these systems have a novelty factor, and get some development bonus. When it comes down to hard numbers, in comparision with other options for a particular application… we will have to see.

• Scalability
When it gets beyond a simple ring or another simple layout, the complexity of the system increases dramatically (keeping in mind that the system is highly distributed intelligence). It may be possible to overcome the problems, but the question remains, at what price.

• Acceptability
Outside of an enclosed application (such as aeroports, amusement parks etc.), the impact of the system’s infrastructure must not be neglected, as it has to be above or below ground level. How well that is accepted in residential areas has to be shown.

• Percieved safety and security
The system is automatic and has small units. This prevents any kind of social control. Outside of enclosed applications, well, maybe even within, people will act … like people… as it happens in public elevators as well. Also, the vehicles run on a rather elevated flimsy roadway, and are rather sealed. This will raise questions

• Maturity
This is more a question on how PRT systems are marketed. At the moment, but that might be with all kind of “unconventional” transit systems, there are lots and lots of buzzwords, but not much substance. Apparently, the publications are aimed at politicians (understandable… they are the ones who decide about budgets). Also, there are claims which are hardly (or very expensive) to prove. This and also the technical reliabiltiy has to be shown.

Quite a bit of homework to do…

>rises from 55% to 57%.

Correction: 55% to 58.5%

@EngineerScotty:

“Yet many of the comments from PRT supporters above–continue to frame PRT as a replacement for mass transit (or at minimum, as more worthy than traditional mass transit for the limited funding that’s available); and some continue to promote the myth that rail transit is only suitable for cities like New York or Hong Kong, with five or six figure density figures (persons/square mile).”

Some mass transit is operating where it is not really viable, and is only able to operate with very high subsidy and/or low frequency of service, due to the ridership demand being insufficient to fill the vehicles. In those cases, I would be in favour of PRT replacing mass transit. There are also lots of places where PRT can operate that are currently not served at all by mass transit. These are PRT’s “low hanging fruit”, so I don’t think there is any serious threat in the short term to mass transit jobs from PRT. If vendors of mass transit systems are worried about a threat from PRT to the long-term future of their business, they can always develop their own versions of PRT. There is nothing preventing them. After all, the key patents for PRT guideways, switches, etc., have all been in the public domain for years. I have no affiliation to any PRT vendor.

Bruce,

I’ll agree that “some mass transit” operates where it requires a larger subsidy to provide acceptable level of service (what level of “subsidy” makes a service “viable”, and what LoS is “acceptable” are political decisions, so I’ll punt on these), but there are often good reasons for doing so. Many transit services are run for social benefit, rather than for mass-mobility or environmental reasons; generally these things aren’t light rail.

But that’s a strawman argument: Except for toll roads, streets have a farebox recovery of zero–and even if you count 100% fuel tax revenues as part of the revenue equation for streets, they still require subsidy. Many try and hold transit to a higher standard of financial performance than they do streets.

That said, there are STILL many arguments out there that PRT systems ought to be deployed instead of light rail systems for applications where ridership and/or service quality is expected to be low for a given budget. Perhaps, but again–there are other transit technologies that are neither LRT nor PRT that also fill this niche well.

Let me ask a question: Why, do PRT supporters suppose, production PRT systems have not been implemented other than a handful of cases?

Is it:

* Opposition and FUD from bus and rail interests (manufacturers, transit unions, and others who have a financial stake in other forms of transit)
* Opposition from transit opponents altogether–the auto and petrochemical industry, the political right wing, and/or motorists who like their cars and don’t want to pay for anything else that they don’t plan to use?
* Intransigence and/or excessive risk-aversion among transit planners, politicians, and others responsible for the design and operation of specific systems, who are reluctant to try something new?
* Issues with regulatory authorities?
* Objection from “urbanistas”, some of whom oppose suburbia and low-density development altogether (including for reasons other than environmental concerns), and who may be opposed to any transport mode which caters to the suburbs?
* Objection from NIMBYs, who might dislike PRT infrastructure in their backyard (even those more than willing to tolerate streets)?
* Objection from private developers and their financiers, who prefer to build (or finance) development in forms which are well-known and well-accepted, and unwilling to risk money on something new?

I’m interested in who you think is holding PRT back.

@EngineerScotty:

I’ll agree that “some mass transit” operates where it requires a larger subsidy to provide acceptable level of service. . . Many transit services are run for social benefit

In many cases, it may turn out that greater benefit can be obtained at lower cost by PRT than by bus or rail.

Except for toll roads, streets have a farebox recovery of zero–and even if you count 100% fuel tax revenues as part of the revenue equation for streets, they still require subsidy.

You are doing that special pleading thing again. Rail is much more heavily dependent on subsidy than roads. How much of the road system is paid for by the drivers varies from country to country, but in the UK where I am, around 60-80% of a driver’s fuel bill (depending on current prices of fuel) is a duty paid to the state. Add to that annual Vehicle Excise Duty, which can be up to £400 a year for a car, Value Added Tax on cars at 17.5%, and compulsory insurance policies, also taxed, and it turns out that drivers and car owners, rather than the state (as routinely but inaccurately claimed by rail advocates), pay for the road system.

Many try and hold transit to a higher standard of financial performance than they do streets.

And many, evidently, do not.

Let me ask a question: Why, do PRT supporters suppose, production PRT systems have not been implemented other than a handful of cases?

I don’t think there’s any big problem on that score. It costs millions and takes years to get a test track developed, so we should expect that stage of development to be fairly slow. The first successful PRT test track (Cabintaxi) was built 30 years ago. In the end, mere luck (in the form of an inconveniently-timed financial crisis hitting the German government) prevented its deployment. The second one (Raytheon) was a flop, many obvious errors having been made. The third one (ULTra) was built in 2002 and was a success, leading to the first commercial sale. That’s not too bad, compared to other innovatory technologies that involve high initial development and start-up costs. Some ideas with apparent promise never get that far. Others are tested for years without ever getting to a sale.

I’m interested in who you think is holding PRT back.

Very little. It’s progressing healthily.

I would think mode share divided by overall investment would be a better metric

If we take investment into account, rail looks even worse. Rail was the dominant mode of transport until the mid 20th century. Since then, it has sucked in countless billions in subsidy, while delivering an ever-diminishing mode share. Meanwhile, car drivers pay for the roads through various taxes and duties. In the UK, car drivers pay approximately £40 billion a year in fuel duty, vehicle excise duty, etc., but the government only spends £10 billion on the roads. In other words, car drivers not only pay for the roads in full, but they also subsidize the railways, despite that the railways only provide 1/15th as much service (measured in passenger-km) as cars.

until transit can somehow pull itself up by its bootstraps and boost its “share” ex nihilo, which may in fact happen if and when gas hits $5/gallon in the US.

Unlikely. The preference for driving over train travel is impressively resistant to price. Petrol is $8 a gallon in the UK (due mainly to fuel duty), and yet the mode share for cars is 87.8%.

A more pernicious form of the argument is that transit is therefore a “failure” and should be defunded.

I think it’s a good argument. Let all systems duke it out on the market. My bet is that if they do, trains will soon be extinct, but buses will probably survive in a few niches, and PRT will thrive.

if the funding is diverted from auto infrastructure.

You repeatedly suggest that the auto infrastructure survives on subsidy. It does not. Indeed, as we have seen, in countries (such as the UK) where the auto infrastructure, far from being subsidized, is treated by the government as a cash cow, the car remains the overwhelmingly dominant mode of transport.

Just because 60% of trips (according to Bruce) in Japan are taken by autos

Not according to me, but according to Japan’s official statistics:

http://www.stat.go.jp/english/data/handbook/c09cont.htm

As for cars, you’re only counting direct spending on roads, not spending on pollution reduction, carbon offsets, and all the other environmental externalities of roads. The fuel taxes offset the externalities, partially; Greg Mankiw has argued that driving generates $2.21 in pollution externalities per gallon, and another $1.00 in carbon externalities. No country sets its fuel taxes significantly higher than the total, which doesn’t even include military protection of oil sources…

The idea that all fuel taxes should go to direct road spending only is a uniquely American idea, peddled mostly by motorist clubs and rural populists going back to the 1910s. Britain instead went for taxes discouraging pollution. So in Britain, roads at least pay direct costs, if not indirect ones. In the US, they don’t; the state of Texas has done an audit and found that not a single road in Texas pays more than half its construction cost, and some pay 16%.

See this link for a more exhaustive list of road-related costs borne by the general public.

I am a transit user. I have lived in and used transit as my primary means of transport in Pittsburgh, PA, Wash DC, Atlanta, GA, Seattle, WA, and San Dego, CA. I am a PRT advocate because I want better transit.

Buses are the predominent form of public transit in the US. Unless a bus system is so small as to not be able to justify its own fuel depot, bus systems do not generally pay fuel taxes. In other words, they have free use of the majority of their infrastructure. Even with free infrastructure these systems can not recover O&M – not mention capital costs – from the fare box. The reason for this is twofold. Drivers are expensive and running buses empty is expensive.

PRT suffers from neither of these issues. Unfortunately, PRT infrastructure is not free. While nowhere near as expensive as LRT infrastructure, PRT infrastructure is not cheap.

I am constantly amazed and amused by opponents arguements against short headways. In the US, we accept as SOP headways of less than 2 seconds on every freeway in the country. Headways performed by irrascible, incompetent, highly variable humans on a “guideway” with unrestricted lateral movement. Yet having the same operation performed by fault tolerant computer on a resticted grade separated guideway – where the worst that can happen is to rear end the vehicle in front of you – is deemed to be impossible.

I am irritated by transit “advocates” who do not use transit themselves. If it’s good enough for me, it damn well better be good enough for you! I still remember my first transit meeting in Seattle. I got up and asked how many people arrived at the meeting by transit. 3 people – including myself – raised their hands. Who the hell were the rest of these people to be telling me what my transit should be?

Michael Setty is a “transit consultant” who makes his living pushing buses and LRT. He has been invited repeatedly to join the Transit Innovators (TI) listserve on Googlegroups (easily accessed via http://www.ATRA.org). On this listserve we have people who are engineers, economists, computer programers, and system vendors. At least two of these system vendors are former aerospace engineers. We are still waiting for Mr Setty to show up.

The tax on a litre of petrol at the pump in the UK is currently £0.5419 fuel duty, plus VAT at 17.5%.

VAT is a general tax, not just a fuel tax. All goods and services are subject to it. It’s just in the US that gas is considered so special it gets an exemption from sales taxes.

pstransit:
On this listserve [Google Transport-Innovators group) we have people who are engineers, economists, computer programers, and system vendors. At least two of these system vendors are former aerospace engineers. We are still waiting for Mr Setty to show up.

But you have no one who has ever actually worked “hands on” in public transit, particularly at the operations level as I have for 30 years, as far as I can tell. You certainly would have mentioned such individuals–if you had any on the T-I list.

I am not going to join the “Transport Innovators” Google group because it is a waste of time arguing with PRT true believers, who have the gumption to lecture me about what I actually or should know about transit operations “on the ground.”

I don’t presume to lecture aerospace engineers about the numerous details and “tricks of the trade” in their profession, nor do I lecture software programmers about the intricate details and various hacks and tricks required to be successful at what they do for a living.

But I’ve been the subject of name-calling, accused of getting paid by the “light rail industrial complex” and not knowing anything about public transit planning or operations, despite having written dozens of transit plans over my career, as well as dealing with numerous transit operating, finance, management and institutional issues over the years. As well as lectured by a software programmer who clearly understands the complexity of AI, but shows little acceptance of the myriad complexities, particularly human, of operating transit, PRT or otherwise, in the real world.

Reasonable people will understand if I don’t bother with such an errant waste of time.

BTW, I have more than two papers or presentations by PRT advocates that, interpreted correctly, shows that Vuchic’s point about the capacity and economics of PRT are absolutely correct. I’ll post something on this to my website very soon, after I get some real work done–so it may be a few days.

For once I agree with Peter Mueller, to a point.

There is indeed much to learn from Morgantown. The most important lesson is the need for much larger vehicles than 4-person podcars, if an automated guideway system is to operate effectively and efficiently with sufficient capacity. This includes line capacity andproviding sufficient station throughput to meet peak demands, at least in markets where there is enough potential transit traffic density to justify construction of expensive fixed guideways in the first place. Again, Morgantown is a case study proving many of the points made by Vuchic a long time ago.

To Max Wyss:

I think the size of bigger vehicles would depend on the likely traffic volumes that a corridor is projected to produce. While Morgantown runs with 8-seat, 12-standing vehicles running at 15-second headways (240 cars per hour at maximum, with up to 4,800 passengers capacity per hour, in principle), either an LRT or more conventional automated guideway with significantly larger vehicles in the VAL range or larger, would have been simpler, and probably somewhat cheaper to run, allowing West Virginia University to operate service during low demand times when school isn’t in session, like on Sundays.

The Morgantown system works for what it is. Rebuilding it with the same technology probably would be cheaper than tearing out the existing system, if only because of the massive sunk investment in the system’s massive infrastructure. In fact, I think this is the plan that university and city decision-makers are pursuing with the Federal government.

You pay VAT if you drive to work, but not if you take the bus or train.

So? Britain has decided to exempt goods and services used by the poor.

Now, if you’re talking about Calgary, which started planning in the 1960s by setting aside ROW for future rail lines, then rail may be more competitive. But, in North America, Calgary is the exception and most cities are not positioned whatsoever for rail.

Calgary’s ROW reservation did two things. First, it limited construction cost. That would be true for every mode of transportation, whether rail, BRT, or PRT. And second, it made the main routes known in advance, so the city could upzone them and redesign buses to feed light rail; again, without such plans PRT would not succeed, either, because uniform low density would be too expensive to serve by a $14-30 million/mile system.

But Calgary is not the only city where light rail demand far exceeds the capacity of PRT – at most, it’s the only such city in the North American Sunbelt. This demonstrates competence, which is sorely lacking in the rest of the North American transit planning world. The problem is that you can’t argue for PRT on the grounds that LRT is designed incompetently, because the same urban planners would be doing the system design.

As if that weren’t enough, a certain Professor Smeed at the University of London came up with the idea sixty years ago that expanding the road system to accommodate cars would be a bad idea, since it would merely lead in the long run to increased car use, and the expansion of car use would cancel out any relief to congestion that the new roads or new lanes were intended to provide. Governments in the UK ever since have been influenced by this idea, and have as a matter of policy strangulated the road system. Many towns in Britain have pedestrianized all or part of their central area, banishing the car outright, and London was one of the first cities in the world to introduce congestion charging.

This is such a load of crap I don’t even know where to begin. In the 1960s, after Smeed released his report concluding that expanding road capacity would not reduce congestion, Britain studied his report to death and did nothing. The only country that listened was Singapore, which was technocratic enough for the government to be able to implement such things without opposition. Singapore’s road pricing program was a success, but even then, the West didn’t try to learn until decades later, when Singapore became rich and globalized and Westerners started to visit. London implemented congestion pricing in 2003, 28 years after Singapore. Calling it “one of the first cities in the world to introduce congestion pricing” is true but misleading.

Most Britons do not live in rural areas, but most Britons do not live in dense cities, either (i.e. cities where it makes sense to spend $14-30 million/mile on fixed-guide transportation). The idea that they still live in medieval city boundaries is laughable. The medieval boundary of London is the City of London. The late 19th century boundary is most of Greater London, which itself is barely half the London metropolitan area. Most of the rest was built for cars: the 19th century cities in Northern England were depopulated after the war because of urban renewal schemes. Where it wasn’t built for cars, cars don’t predominate: in Greater London, transit has a 52% mode share, private vehicles 37% (see table 1.8). And while the transit share has risen in response to congestion pricing, there was an equally large increase, driven mainly by a reduction in car traffic, in the years before 2003 (see table 1.6 in the above link).

PRT is the worst transit “solution” and its backers are delusional, obsessive, and quite frankly, not pleasant to be around. They are the most pathetic in a sea of already pathetic people, transit advocates.

Mike C, I’m sure you don’t live in your mother’s basement.

“Peaking” of 15% is for the dominant direction; this percentage is higher when you include the non-dominant direction.

FWIW, one method to shorten station dwelling time is using insular tracks with platforms on both sides; one for getting off, and the other for getting on the train. Such a layout can be found at the busiest stops of the Munich S-Bahn tunnel (at Hauptbahnhof and Marienplatz). I don’t know whether BART has a similar station layout. Another (more expensive) approach would be doubling the tracks at a station.

The paper mentioned is quite interesting, and it does put the capacities in serious relationship. The capacities mentioned, based on 3, 2 or 1 second headways can only be reached on “trunk sections”, and require a very high number of vehicles.

I can not prove it (as my math is a bit rusty), but it looks to me from reading the paper, that the capacity could be somewhat increased by pre-assigning berths and queues to specific destinations, instead of letting the passengers specify the destination when they get on the vehicle.

I have my doubts whether such queue management would work unattended.

To Peter Muller, message 205:

With such a 15-berth station, the capacity limit is indeed not the boarding zone (assuming that a vehicle every 20 seconds is practically feasible). The bottleneck (as well as the challenge for the control system) of such a station is the entrance to the main guideway, where you have to get the vehicles up and running at a 1.25 seconds headway. And, in order to provide access to every destination from every berth, you will need a section through which all vehicles pass. We now could use some simple physics, and guesstimate how big that accelleration zone would become…

Also, for playing around a bit more with numbers, assuming that a vehicle would be back at that 15-berth station within 6 minutes, you would at least need 270 vehicles.

Such a system is already quite a monster, IMHO. And I would not be surprised if it would break down under its own weight… one had to do some simulations.

Max Wyss:
I can not prove it (as my math is a bit rusty), but it looks to me from reading the paper, that the capacity could be somewhat increased by pre-assigning berths and queues to specific destinations, instead of letting the passengers specify the destination when they get on the vehicle.

You are correct. In other words, capacity could be significantly increased by operating in a semi-scheduled mode, and even more if each berth served a scheduled stopping service with larger vehicles, as shown by Morgantown and every real transit system. Your point shows the utter futility of PRT in coping with all but very tiny passenger volumes, at levels where its capital costs cannot possibly be justified by any rational metric. Of course, this argument and thread has hardly been rational.

Mike C., an elevated PRT system with ULTra style berths will be somewhat wider than most designs with in-line berthing, so is unlikely to be constructed due to its higher cost. And the “little old lady” problem” impacts the STATION CAPACITY issue, not line capacity per se. Any transit system, PRT or otherwise, has to have some room for error. Why you don’t get this is beyond me. I will never concede this point, and you’ll never convince me otherwise.

Mueller:
A PRT station could have 15 or more bays for a capacity of 2,700 people/T-Pods per hour so stations are not the [THEORETICAL] bottlenecks you think they are.

Let’s see, 15 bays with ULTra-sized vehicles would be 300+ feet long, comparable in size and cost to an elevated HRT or LRT station. The ULTra docking bays aren’t practical for elevated stations, and add a few seconds at least to each bay’s throughput.

Mueller, you are the one who has completely misunderstood Lees-Miller’s paper.

Part of their simulation included estimates about how long the waiting queue would be at PRT stations, and how this waiting could be reduced through ridesharing. The simulation also allowed for the PRT system to deliver empty pods to the theoretical station. Figure 2 refers to the average wait per “party” whether the party is 1, 2, or 3 persons. In the case of a network with 25 destinations (e.g., my Winona PRT thought experiment), the average wait with 114 “parties” per hour (114, 228, or 342 persons) is 3 minutes; with 216 “parties” per hour and a 25 destination network, average wait of 6 minutes.

Certainly more bays would reduce the wait time for PRT at a given station, as shown in the 1978 PRT book, but in the real world station size and cost has to be balanced with likely demand and the available budget.

You’d also have to have far more accurate patronage modeling than the current state-of-the-art, which has gotten better at system level projections, but still can be highly inaccurate in station-level projections. When LRT station projections are off, the system can usually cope; if you’re low by one or two bays at a proposed PRT station, you’ll have major operational problems.

Projected PRT capacities at stations always break down because (1) you’re not going to get more than about 1.2+/- persons per party in a real world urban setting; and (2) station sizes are strictly limited; 25 elevated stations in Winona with 8-15 bays each would easily cost $3-$4 million each, putting the capital costs of PRT relative to likely capacity even higher in the stratosphere.

When “climbing the steps” or “ride the elevator,” and ticketing and destination selection is included, average total waiting time is LONGER than average waits of buses or trains running every 5 to 10 minutes, at ANY PRT station with more than a handful of passengers at a time. If this weren’t the case, Lees-Miller would also not have simulated QUEUES in the Heathrow ULTra stations, which are greatly reduced by ridesharing–as was the projected number of vehicles from 47 down to 18.

So, in the real world, Vuchic’s bottom point about PRT is made, again: WHY BOTHER with PRT beyond the tiny capacity it can offer, particularly at stations?

To continue playing numbers with our 15-berth station, it would actually be possible to build it on ground, but the access to the berths would have to be from an upper level, going down something like 2 m (how, whether stairs, elevators or escalators would only affect the cost, but not the capacity). Hmmm… elevators… lifting up the vehicles and putting them down again… another interesting idea, which would allow getting on and off from the same level, but would probably seriously extend the 20 seconds turnover time.

Anyway, the actual passenger-related area of such a station would be about 10 x 45 meters on level 1, including the passenger handling space.

However, on ground level, the station would be something like 250 x 50 meters, in a parallelogram shape.

As a comparision: a light rail stop is about 10 x 45 m (serving trains up to 45 m in length (which translates to some 100 seats plus another 250 standees), 6 m width for the right of way, 2 m each side for the platform (or 4 m for an insular platform). Such a station easily serves a train per minute (exactly, one train per minute and direction).

Now think how you would use up the 250 x 50 m for a transit station…

“low density” comes to my mind.

Jarrett:
Only 210 comments! I used to feel flattered by long comment chains too, until I realised that most of them were debates that wandered far from the substance of the article.

Actually, this is hardly the longest PRT argument threads I’ve seen; if Avidor was posting here, you’d see it well over 300 by now! :-)

Michael Setty, almost every assertion you are making here with apparent authority is completely, utterly wrong. As it happens, I’ve been busy today and haven’t been able to respond quickly, but I am prepared to address and correct every one of your recent misconceptions on PRT operation.

Let’s start with station capacity. Let’s assume 10% LOLP (“little old lady percentage”) and that each “little old lady” causes the cycle time to be 45 seconds rather than 15 seconds. For PRT, this means that for every 9 vehicles which load in 15 seconds, the 10th takes 3 times longer. So in any given berth, 10 vehicles load in 180 seconds, for a 18 seconds average load time, rather than the 15 seconds that would be typical of a purely younger crowd.

That’s a 20% increase, the equivalent to less than a single extra berth in all stations up to 5 berths. So if the PRT is installed in a south Florida retirement community, you may have to add extra berths per station. But in college towns like Winona, the LOLP is probably more on the order of 1-3% – insignificant.

But wait: none of this affects PRT line capacity, particularly not ULTra PRT with its parallel stations: if a vehicle takes longer to load, it sits in the station and doesn’t block the line of traffic moving past, and for ULTra, doesn’t even affect other berths in the same station. In fact, in most cases, the little old lady doesn’t affect other passengers one bit, as they just go to a different berth (most stations will have 2-3 berths).

Only in the very busiest times, when there happens to be a full station and all berths have passengers waiting, will passengers be affected, and their average wait will be 20% longer, 18s for every group ahead of them instead of 15s. Even if the berths each have 20 people waiting, which would only happen during an extreme surge such as end of a sporting event, passengers would have to wait an extra minute. Oh, the humanity!

Now consider when that same 10% LOLP is applied to light rail. Again assuming 1 in every 10 passengers is slow loading, and that most stops during busy periods will have more than 10 people loading, this implies that most of the light rail stops will have to stall a whole train waiting for the little old lady. In this case, all passengers wait for our poor LOL: everyone on the current train, everyone loading, and if the trains are running at high frequency, every train behind the the current train on the line.

So, once again we see an example where the individual, parallel aspects of PRT make it less susceptible to “human factors” than buses or trains. A high percentage of “little old ladies” (10%) only adds a marginal amount of time to PRT station boarding time, and even then only in cases where all berths in a station have lines of waiting passengers. But for trains and buses, LOLs at every stop will add minutes to the total journey time, wreak havoc with schedules, and possibly even hold up following trains coming into the station.

Thank you, Michael, for pointing out yet another way in which PRT service is superior.

Alon,

Yes, other people can board through other doors, but the train must sit and wait for every last LOL to board before leaving the station. I.e., one LOL holds back an entire train/bus for the entire duration of the delayed loading time.

@Michael Setty

The Ridesharing paper:

You are completely misrepresenting the thrust of that paper. That paper describes a potential ridesharing alternative for small networks which would allow them to purchase fewer vehicles to service the same demand. The authors fully acknowledge that ridesharing is not a practical solution for networks with large numbers of destinations, so this is obviously intended as a solution only for very small networks, or as a temporary “bootstrapping” solution for small pilot networks which may grow later…

…like, perhaps, a pilot application for an airport that’s starting with only two destinations but which will likely expand later. Sound familiar?

The initial Heathrow pilot has only 18 vehicles, and may benefit from ride-sharing in busy times. The operative word here is PILOT. This is the first installation of a brand new implementation of a brand new transit system. Do you think maybe they didn’t want to build 100 vehicles right out of the gate, in case that the pilot revealed issues with the initial design?

My guess (and it’s purely a guess, but it makes sense) is that ATS proposed ridesharing in order to limit the number of vehicles required for the pilot, and then if minor vehicle design problems were revealed during the pilot, the problems could be fixed in rev 2 and only 18 vehicles would need to be retrofitted. This would be entirely in line with ATS’s approach from day one: walk before running, and crawl before walking.

Having fewer vehicles for the pilot would also minimize the financial risk for the pilot customer (BAA), as the cost of ~40 vehicles would be saved should the system prove to be a failure.

And ridesharing would work well for the Heathrow pilot, because it has only 3 destinations.

Did I mention that Heathrow is a pilot system?

Now let’s consider future systems. As the number of destinations expands, ridesharing becomes increasingly less effective, as the probability that random unrelated passengers will share destinations approaches zero. It is then that the number of vehicles will need to increase to cover the demand.

And at that point, vehicles can be added. Assuming the cost of a vehicle is $50,000 (somewhat high, but these are not mass produced — yet), an additional 100 vehicles will cost $5M of additional capital, and that will cover a mile of bidirectional guideway at about 2 seconds headway. So no more than $5M per mile of additional capital per mile gets enough vehicles to saturate the guideway in the current version. And no ridesharing is needed. If PRT becomes popular, over time those vehicle costs will come down as production increases, so the cost becomes even less significant.

I will continue to address Setty’s arguments in the next few days.

Michael D: Can carsharing be used to commute to work?

Car sharing does seem difficult to use for one-way trips, though Autolib’ in Paris will be doing just that. However, it raises a question: can PRT be used to commute to work? Work tends to be concentrated in CBD’s — precisely the kind of area PRT cannot serve. Can PRT serve even one 10 story office building of 9 to 5 workers? 20 story? And don’t tell me that people will want to abandon driving in order to take PRT and then transfer to mass transit (which apparently they hate).

And it certainly seems to me that PRT can’t really be used to go to the kind of big box stores low-density areas have spawned.

@Michael D: yes, I envision PRT serving CBD areas in moderate density cities. It would not, in its current incarnation, serve NYC or anything near that density, except perhaps in a transit circulator role in outlying areas.

But there are enough moderate density cities out there where PRT can play a significant role in the daily commute, especially in North America.

I view carsharing as a complement to transit. It allows people to have the occasional use of a car when they need it, but relieves them of car ownership. They would still require some form of transit for many tasks, mainly the daily commute, and in many cities PRT can fill that role.

So PRT would complement carsharing just like any other public transit system. It doesn’t have to be either-or.

@Alon,

You are missing the point. In PRT, one to four passengers enter a vehicle, sit, the doors close, and the vehicle moves. For ULTra, this happens in every single PRT berth independently, and for all PRT, it happens in every station with no affect on the line.

If one PRT vehicle loads slower than the others, it has no effect on other berths in that station, and only has an effect on other passengers IF there is a queue at that particular berth, and a marginal effect at that.

For a bus, a slow loading patron holds up the entire bus, and likely holds up vehicular traffic behind the bus. Potentially dozens of passengers are delayed by a single LOL, as opposed to the occasional PRT passenger who happens to get caught in a line behind a LOL when there are no other empty berths.

Mike, loading for a car is slower than loading for a bus, per passenger. But the issue is not what happens at one berth; it’s that you need many berths to have high capacity, and those can be blocked easily. The appropriate transit analogy is to ticket vending machines, where 3-4 slowpokes can create long lines for everyone.

Also, Alon, how can you claim that loading 1-4 passengers into a single vehicle on a level platform is slower than loading perhaps half a dozen or more people, all through a single narrow door, and up 2-3 steps? And we all know what effect the steps have on our “little old lady” or someone in a wheelchair, both of whom would walk/glide right onto a PRT vehicle with no steps to traverse.

Boardings involving steps do pose problems for the disabled, or for LOLs.

Many modern trains, including LRT, involve wide doors and level boarding, which reduces the affect of a LOL or wheelchair.

Local bus service, where everyone must file past the driver, is affected much more by LOLs, but that’s hardly rapid transit…

Oops, mispelled someone’s name: Setty, not Shetty.

You have a cheek accusing me of spreading disinformation, when what I am doing is correcting your disinformation.

Most Britons live in towns and cities that are denser than US cities. The population is about 90% urbanized, and the average density of urban areas is considerably higher than that in the US. Street layouts date back mainly to the era before mass car ownership. All this is quite contrary to the impression that you tried to convey. Urban renewal in the North did not cause depopulation as you suggested, nor was it car oriented, but in fact often involved pedestrianization and/or upgrading of public transport.

You tried to give the impression that the car is favoured by government policy and expenditure, etc., in Britain, relative to public transport, and that most urban areas in the UK are designed around the car. Practically everything you said about driving, roads and urban planning in the UK is the directly contradicted by the facts. Simply, you were spreading disinformation, and I corrected you.

We’re talking about the country that has the least motorway per capita among the EU15 (less than half as much as most other countries in that group), the country that for several years imposed an nationwide moratorium on road-building, the country that has the highest fuel duties in the whole EU, the country all of whose largest cities acquired most of their present size before the majority of people owned cars, and thus whose street-plans are not designed around the car, the country where freeways are conspicuous by their absence.

It is simply crazy to suggest that the above-described country is a country where people drive because driving is subsidized and/or encouraged by the state. You engage in blatant cherry-picking of data when you take a country where nearly 90% of km traveled are by car, as opposed to around 6% by train, and try to convey the impression that the train is more popular than the car by selecting an outlying single datum that would seem (misleadingly) to support that, namely, trips to work in London. Right next to table 1.6 in the document you cited is table 1.5, which shows that, while public transport’s share of commuter traffic in the UK is 22%, its share of non-commuter traffic is a third less, at just 14% (and even this figure is biased in favour of public transport, because of how it represents mixed-mode trips). So trips to work do not tell the whole picture.

All your disinformation is aimed at supporting the false notion that mass transit is the ideal form of ground passenger transport, and people have no rational grounds for preferring other modes. The truth, which you persistently evade, is that, whatever the downsides of the car regarding, e.g., pollution and land-use issues, travel by car is usually faster, more convenient, more flexible and more comfortable than travel by mass transit, and these quite rational considerations, rather than any conspiracy by governments or corporations to favour the car, explain the car’s persistent popularity. It is not feasible to upgrade mass transit the the point where it is as attractive as the car, therefore, without authoritarian measures or some kind of spectacular economic collapse or extreme fuel shortage, mass transit will never dominate ground passenger transport. If you want to support a public transport system that does have the possibility of attracting a high proportion of ridership away from cars, then PRT is far more promising than trains and buses, whose relative unattractiveness to riders is proven.

On the ULTRA website, it is written that the protection sytem is based on a fixed block system inside the guideway. Does anyone has information on that ? I can’t understand that finally they don’t use use only the distance to the leader inside the pod to provide safety. This is totally in contradiction with the PRT goals … and it totally mimick the railway system. The only thing that remains is thus the on demand. I suppose taht they were unable too proce the safety of the pod independantly thus they had to rely on another system.

They don’t even use moving blocks thus the pod speed won’t be taken into account in pod distance. Maybe this is the reason for the delay because the safety was not considered sufficient ? If the headway is increased too much, the detection system inside the pod has not enough range to detect the leader pod or is unable to see through the curve and grade. Thus you must have an alternate system to see further and keep a sufficient distance ahead.