I'm wondering about that throughput estimate. I *love* the idea of gondolas, which would soar above so many of our current problems. But with existing gondolas, throughput seems to be a problem. The claim here is 3000 passengers/hour. That would mean almost 1 passenger per second. I guess I could imagine a cable with 1 pod screaming past every second (or one pod with 5 people every 5 seconds). But the logistics of loading and unloading all those different pods at the ends seems really challenging. It seems like you might need giant stations with space for lots of pods, a big "runway" for the pods to reach ambient speed, a really clever merging algorithm, pods spaced very close together, and everything running independently with near 100% reliability. This doesn't seem impossible, but it seems really hard.
It does seem tricky. I suspect part of the answer is a feature you didn't include in your list, which is 'many stations'; it's hard to get to 3K/hr on (say) five stations to handle the loading and unloading, but it would be easy on 50. If I'm right, this level of throughput is feasible, but only for a mature and elaborated system, not a nascent one.
You're asking the right questions. Like with Waymos, the car is an incredible piece of hardware, but it's the software that makes it all happen. We are assuming vehicle spacing of around 2.5 seconds. Transit normally uses the full vehicle size when calculating capacity, but are more conservative. And Andrew is right, that it takes a few stations to spread the load.
Unlike conventional gondolas, our stations are offline (or some people call them “pull over” stations. Vehicles leave the mainline cable and come to a complete halt at stations. Meanwhile, other vehicles on the mainline bypass the station without stopping or slowing. So vehicle loading/unloading times in-station have no effect on mainline speeds. When vehicles depart, they merge back on to the mainline cable at speed. So all vehicles on the mainline at all traveling at speed. In addition, control system software allows minimal headways, so that vehicles can be spaced closely together. And while conventional gondolas have single origin and terminus, our system creates a network of point-to-point travel between any two points anywhere in the network.
How does the system cope with peaks? For example, you put a Swyft station in the Theater District. A performance is over and at least 100 distinct groups of passengers decide they want to ride Swyft home.
1. How quickly can 100 gondolas come to the station?
2. How quickly can 100 groups each find their own gondola and embark?
3. Can a station fit 100 gondolas? If not, where will the rest of them wait?
I know it's a last mile system not designed to replace all rapid transit, but the same problem still applies to the interchange stations. What if Houston repurposes the rail line running through Sugar Land east to Fannin South and hundreds/thousands of commuters start to disembark from each train during the rush hour?
Getting skyscrapers to pay for their own station seems eminently doable. Rents per square foot per year for office space next to a major station are likely at least 10 dollars higher than those a mile away. Let's say access to the PUPPET system is worth 5 dollars. A typical skyscraper has 500,000 square feet or rentable space. Then building a station at 10 million dollars per mile will pay for itself in 4 years.
It would be stupid when doing any major development, especially a group of skyscrapers, not to pay for a station.
I love the idea of aerial lifts as a supplemental transportation form. The Randalls Island tram is in my backyard and I’m a huge fan. I think solutions like skylift have real validity but I would be wary of making them PUPPETs. In many ways it would defeat the great benefits of the aerials. We might ultimately end up duplicating the roadways below. You also lose out on the economy of scale that is essential to public transport. The fixed cars provide one of the best benefits of the aerials; centralized motor/power and control. Powering every unit adds costs, complexity and maintenance concerns. The beauty of the aerials are their simplicity!!
Separately, I completely agree that transit routes need to make money. This is not about profit but proper allocations of our public resources. Our transit should serve the demand and provide the most value to the riders!
You're right that an aerial multi-point network loses economy of scale while adding complexity. Of course, it also gains resiliency and redundancy. The alternative lacks these: one failure in the fixed-guideway motivator shuts down the entire line.
The choice of technology revolves around what problem you're trying to solve. If that problem is linking A to B, no branching or extension needed or ever expected, in a situation where roads and bridges are impractical, then a straight-up gondola would be best. Tourist applications in mountains do seem to be the likeliest case.
I've always liked the notion of gondola-like aerial transit, for many of the reasons you have cited. I'd be interested in hearing about the noise situation; a ski lift type of operation (with the car clamped on to the cable) is pretty quiet except at the towers, whereas I would expect an annoying hum of high-speed-small-wheels-on-twisted-steel-cable throughout this system. Probably no more than road traffic, but if it were at the second and third floor level it could be annoying to neighbours. And the visual impact of the towers would also be an issue with the public in some corridors (unlike the image you've used, which shows cables magically hanging from a single overhead gantry).
Like many of these tech-focused transit solutions, a major issue with procurement and municipal implementation is its unique and commercially proprietary nature. Municipalities know how to build and maintain roads and bus systems, and they don't need some high-tech company from California to do that. They can have a competitive consulting process to create the standard design using off-the-shelf components, hire any number of local construction firms to build and maintain the infrastructure, buy vehicles through competitive bidding, and pay (unionized) staff to operate things. Much the same goes for railways. Governments are terrified of getting locked into some (foreign) private company's proprietary system, where they have little control over incident response, spare parts, operating knowledge, and the operating company's financial stability and longevity. And unions hate this stuff because driverless systems "take away" operating jobs. You need a visionary government to commit to something like this (e.g. Ontario with Skytrain) and even then, the potential for success seems limited, because in any different market you need another government to have the same willingness to leap into the void. And so on and so on in every market, until it becomes a standard option for everywhere. Just look at how long it has taken for congestion charging to become "normal" (i.e. decades, and not yet) despite its compelling merits and proven implementability.
Governments are also obsessed with having competitive bids and demonstrating "best value" to their taxpayers, so they have no procurement mechanism to handle a unique or non-competitive piece of infrastructure, particularly a big-ticket city-shaping item. When I worked with the City of Toronto, we were approached occasionally with something like this, and we simply could not handle it on the procurement side because there was no competitor to choose from.
And the PRT (Personal Rapid Transit) aspect of it is most appealing; it is kind of amazing that so little progress has been made on that front over the decades (hence Uber, Lyft, and taxis picking up the slack). You may know that a cool but relatively unknown rail-based quasi-PRT system has been operating - in West Virginia, of all places - for decades: https://en.wikipedia.org/wiki/Morgantown_Personal_Rapid_Transit
Your points are well taken, Steve... as you know, the whole Sidewalk Labs misadventure came to grief in large part because of municipal inflexibility. Against this, while roads-and-buses are open non-proprietary technologies, rail of any sort is not; there is a small market but not a very competitive one, which is one reason for LRT debacles we saw in Ottawa and we're seeing now in Toronto. Relative to LRT or heavy rail, Swyft seems to be a lower-cost, lower-risk approach, which I hope gives them runway to build up markets and establish credibility.
I wonder if these hurdles are easier if the PUPPET is private built, owned and operated, but with clear up front obligations to the municipality. Then it becomes a free win for the municipality - they get a new transport solution at no cost.
Has this technology ever been considered for use in the Bay Area? It might avoid some of the construction difficulties that have hampered MUNI expansion (especially on the west and north side, where rail has yet to reach). I know that the SF Chronicle reported in October that it was investigating gondolas as a potential way of getting people from Forest Hills station to the hospital, and I didn't know if that was a competitive bid process.
Additionally, does the Swyft City technology lose the comparative benefit of gondolas re: hills? My understanding is that moving cable versions avoid traction / energy costs when going up steep hills, but a self-propelled car version along a fixed line would seemingly be similar to a wheeled vehicle. Also, are they earthquake-vulnerable?
1) I suppose it depends on your definition of 'Bay Area'; Swyft got its start as Google's 'Project Swyft' for their Mountain View campus. I'm not sure, beyond that
2) You are right that moving cable versions would have energy savings that moving vehicle versions of gondola tech would not. It's all relative, of course; the Whoosh vehicle would require much less energy than an electric shuttle or bus
3) Regarding earthquakes, I can't speak for Swyft on this point, but I expect the cables aren't perfectly taut, so they would be fine in minor earthquakes. Earthquakes powerful enough to fell one of the support posts would be a different matter.
I'm interested in how the user experience/software for this is meant to work.
I'm hoping for something like. You select a target station in the app and it gives you your gondola number. That gondola will plot a course for your target station, stopping to pick up other people who can share the same route on the way.
Pretty much. But normally you wouldn't expect it to pick up others along the route. In any autonomous vehicle, there is a danger from adding an unknown passenger to a vehicle with someone already inside. So carpooling would happen at the origin (eg, train station), and could drop people off along the way, but you wouldn't expect people to be added along your trip.
Perhaps the problem can be solved if we just have the first rider who volunteers drive the bus until they get to their destination. This would basically cut the cost of bus service in half. Car rental companies already make money renting their cars to random people (of dubious and uncertain driving ability) and how hard can driving a bus really be?
Sort of the Lime scooter model of public transport.
It takes 4 weeks to learn to drive a bus safely, how to deal w/ difficult passengers, what to do in emergencies, how to drive w/ air brakes - a separate driving licence rating. Most thorough training I’ve ever done. Driving such a wide vehicle on crowded streets w/ pedestrians about, pulling in & out of bus terminals, sometimes underground, is often treacherous. I lasted 4 weeks in revenue service. I would not stay on a bus driven by someone untrained.
Generally, no; they charge tenants enough in rent to more than offset the cost of elevator operation and maintenance (and the capital debt to install it in the first place). The analogy does not readily extend itself to transit... the elevators only benefit the occupants of the building, so it's obvious who should pay for it. Transit systems benefit far more people, and far more diffusely, which is one reason the Endless Emergency persists.
Transit systems benefit everyone (even the car drivers). There is no fundamental difference between a tax on residents and leases for office space. Transit systems easily benefit millions of persons whether they ride or not. That's a lot of revenue.
Sure, transit systems *indirectly* benefit everyone in a city, just as a thriving business on the eighth floor of a building indirectly benefits everyone by generating economic activity. But the business can't charge for its indirect benefits; it offsets its costs by charging direct beneficiaries.
It's much more direct then that. 100 commuters in a subway train means 95 fewer cars on the highways.
Luxembourg can provide free transportation for every person in the in the country, but you don't think this can work in the US for a city. You may be quite right that it won't work in the US, but that's a damn shame.
I'm wondering about that throughput estimate. I *love* the idea of gondolas, which would soar above so many of our current problems. But with existing gondolas, throughput seems to be a problem. The claim here is 3000 passengers/hour. That would mean almost 1 passenger per second. I guess I could imagine a cable with 1 pod screaming past every second (or one pod with 5 people every 5 seconds). But the logistics of loading and unloading all those different pods at the ends seems really challenging. It seems like you might need giant stations with space for lots of pods, a big "runway" for the pods to reach ambient speed, a really clever merging algorithm, pods spaced very close together, and everything running independently with near 100% reliability. This doesn't seem impossible, but it seems really hard.
It does seem tricky. I suspect part of the answer is a feature you didn't include in your list, which is 'many stations'; it's hard to get to 3K/hr on (say) five stations to handle the loading and unloading, but it would be easy on 50. If I'm right, this level of throughput is feasible, but only for a mature and elaborated system, not a nascent one.
You're asking the right questions. Like with Waymos, the car is an incredible piece of hardware, but it's the software that makes it all happen. We are assuming vehicle spacing of around 2.5 seconds. Transit normally uses the full vehicle size when calculating capacity, but are more conservative. And Andrew is right, that it takes a few stations to spread the load.
- Jeral (CEO Swyft Cities)
Thanks for the reply, I wish you success!
Unlike conventional gondolas, our stations are offline (or some people call them “pull over” stations. Vehicles leave the mainline cable and come to a complete halt at stations. Meanwhile, other vehicles on the mainline bypass the station without stopping or slowing. So vehicle loading/unloading times in-station have no effect on mainline speeds. When vehicles depart, they merge back on to the mainline cable at speed. So all vehicles on the mainline at all traveling at speed. In addition, control system software allows minimal headways, so that vehicles can be spaced closely together. And while conventional gondolas have single origin and terminus, our system creates a network of point-to-point travel between any two points anywhere in the network.
How does the system cope with peaks? For example, you put a Swyft station in the Theater District. A performance is over and at least 100 distinct groups of passengers decide they want to ride Swyft home.
1. How quickly can 100 gondolas come to the station?
2. How quickly can 100 groups each find their own gondola and embark?
3. Can a station fit 100 gondolas? If not, where will the rest of them wait?
I know it's a last mile system not designed to replace all rapid transit, but the same problem still applies to the interchange stations. What if Houston repurposes the rail line running through Sugar Land east to Fannin South and hundreds/thousands of commuters start to disembark from each train during the rush hour?
Getting skyscrapers to pay for their own station seems eminently doable. Rents per square foot per year for office space next to a major station are likely at least 10 dollars higher than those a mile away. Let's say access to the PUPPET system is worth 5 dollars. A typical skyscraper has 500,000 square feet or rentable space. Then building a station at 10 million dollars per mile will pay for itself in 4 years.
It would be stupid when doing any major development, especially a group of skyscrapers, not to pay for a station.
I love the idea of aerial lifts as a supplemental transportation form. The Randalls Island tram is in my backyard and I’m a huge fan. I think solutions like skylift have real validity but I would be wary of making them PUPPETs. In many ways it would defeat the great benefits of the aerials. We might ultimately end up duplicating the roadways below. You also lose out on the economy of scale that is essential to public transport. The fixed cars provide one of the best benefits of the aerials; centralized motor/power and control. Powering every unit adds costs, complexity and maintenance concerns. The beauty of the aerials are their simplicity!!
Separately, I completely agree that transit routes need to make money. This is not about profit but proper allocations of our public resources. Our transit should serve the demand and provide the most value to the riders!
You're right that an aerial multi-point network loses economy of scale while adding complexity. Of course, it also gains resiliency and redundancy. The alternative lacks these: one failure in the fixed-guideway motivator shuts down the entire line.
The choice of technology revolves around what problem you're trying to solve. If that problem is linking A to B, no branching or extension needed or ever expected, in a situation where roads and bridges are impractical, then a straight-up gondola would be best. Tourist applications in mountains do seem to be the likeliest case.
I've always liked the notion of gondola-like aerial transit, for many of the reasons you have cited. I'd be interested in hearing about the noise situation; a ski lift type of operation (with the car clamped on to the cable) is pretty quiet except at the towers, whereas I would expect an annoying hum of high-speed-small-wheels-on-twisted-steel-cable throughout this system. Probably no more than road traffic, but if it were at the second and third floor level it could be annoying to neighbours. And the visual impact of the towers would also be an issue with the public in some corridors (unlike the image you've used, which shows cables magically hanging from a single overhead gantry).
Like many of these tech-focused transit solutions, a major issue with procurement and municipal implementation is its unique and commercially proprietary nature. Municipalities know how to build and maintain roads and bus systems, and they don't need some high-tech company from California to do that. They can have a competitive consulting process to create the standard design using off-the-shelf components, hire any number of local construction firms to build and maintain the infrastructure, buy vehicles through competitive bidding, and pay (unionized) staff to operate things. Much the same goes for railways. Governments are terrified of getting locked into some (foreign) private company's proprietary system, where they have little control over incident response, spare parts, operating knowledge, and the operating company's financial stability and longevity. And unions hate this stuff because driverless systems "take away" operating jobs. You need a visionary government to commit to something like this (e.g. Ontario with Skytrain) and even then, the potential for success seems limited, because in any different market you need another government to have the same willingness to leap into the void. And so on and so on in every market, until it becomes a standard option for everywhere. Just look at how long it has taken for congestion charging to become "normal" (i.e. decades, and not yet) despite its compelling merits and proven implementability.
Governments are also obsessed with having competitive bids and demonstrating "best value" to their taxpayers, so they have no procurement mechanism to handle a unique or non-competitive piece of infrastructure, particularly a big-ticket city-shaping item. When I worked with the City of Toronto, we were approached occasionally with something like this, and we simply could not handle it on the procurement side because there was no competitor to choose from.
And the PRT (Personal Rapid Transit) aspect of it is most appealing; it is kind of amazing that so little progress has been made on that front over the decades (hence Uber, Lyft, and taxis picking up the slack). You may know that a cool but relatively unknown rail-based quasi-PRT system has been operating - in West Virginia, of all places - for decades: https://en.wikipedia.org/wiki/Morgantown_Personal_Rapid_Transit
Your points are well taken, Steve... as you know, the whole Sidewalk Labs misadventure came to grief in large part because of municipal inflexibility. Against this, while roads-and-buses are open non-proprietary technologies, rail of any sort is not; there is a small market but not a very competitive one, which is one reason for LRT debacles we saw in Ottawa and we're seeing now in Toronto. Relative to LRT or heavy rail, Swyft seems to be a lower-cost, lower-risk approach, which I hope gives them runway to build up markets and establish credibility.
I wonder if these hurdles are easier if the PUPPET is private built, owned and operated, but with clear up front obligations to the municipality. Then it becomes a free win for the municipality - they get a new transport solution at no cost.
Has this technology ever been considered for use in the Bay Area? It might avoid some of the construction difficulties that have hampered MUNI expansion (especially on the west and north side, where rail has yet to reach). I know that the SF Chronicle reported in October that it was investigating gondolas as a potential way of getting people from Forest Hills station to the hospital, and I didn't know if that was a competitive bid process.
Additionally, does the Swyft City technology lose the comparative benefit of gondolas re: hills? My understanding is that moving cable versions avoid traction / energy costs when going up steep hills, but a self-propelled car version along a fixed line would seemingly be similar to a wheeled vehicle. Also, are they earthquake-vulnerable?
1) I suppose it depends on your definition of 'Bay Area'; Swyft got its start as Google's 'Project Swyft' for their Mountain View campus. I'm not sure, beyond that
2) You are right that moving cable versions would have energy savings that moving vehicle versions of gondola tech would not. It's all relative, of course; the Whoosh vehicle would require much less energy than an electric shuttle or bus
3) Regarding earthquakes, I can't speak for Swyft on this point, but I expect the cables aren't perfectly taut, so they would be fine in minor earthquakes. Earthquakes powerful enough to fell one of the support posts would be a different matter.
https://en.wikipedia.org/wiki/Omission_of_New_Zealand_from_maps
I'm interested in how the user experience/software for this is meant to work.
I'm hoping for something like. You select a target station in the app and it gives you your gondola number. That gondola will plot a course for your target station, stopping to pick up other people who can share the same route on the way.
Pretty much. But normally you wouldn't expect it to pick up others along the route. In any autonomous vehicle, there is a danger from adding an unknown passenger to a vehicle with someone already inside. So carpooling would happen at the origin (eg, train station), and could drop people off along the way, but you wouldn't expect people to be added along your trip.
Perhaps the problem can be solved if we just have the first rider who volunteers drive the bus until they get to their destination. This would basically cut the cost of bus service in half. Car rental companies already make money renting their cars to random people (of dubious and uncertain driving ability) and how hard can driving a bus really be?
Sort of the Lime scooter model of public transport.
It takes 4 weeks to learn to drive a bus safely, how to deal w/ difficult passengers, what to do in emergencies, how to drive w/ air brakes - a separate driving licence rating. Most thorough training I’ve ever done. Driving such a wide vehicle on crowded streets w/ pedestrians about, pulling in & out of bus terminals, sometimes underground, is often treacherous. I lasted 4 weeks in revenue service. I would not stay on a bus driven by someone untrained.
Do skyscrapers lose money on every elevator ride?
Generally, no; they charge tenants enough in rent to more than offset the cost of elevator operation and maintenance (and the capital debt to install it in the first place). The analogy does not readily extend itself to transit... the elevators only benefit the occupants of the building, so it's obvious who should pay for it. Transit systems benefit far more people, and far more diffusely, which is one reason the Endless Emergency persists.
Transit systems benefit everyone (even the car drivers). There is no fundamental difference between a tax on residents and leases for office space. Transit systems easily benefit millions of persons whether they ride or not. That's a lot of revenue.
Sure, transit systems *indirectly* benefit everyone in a city, just as a thriving business on the eighth floor of a building indirectly benefits everyone by generating economic activity. But the business can't charge for its indirect benefits; it offsets its costs by charging direct beneficiaries.
It's much more direct then that. 100 commuters in a subway train means 95 fewer cars on the highways.
Luxembourg can provide free transportation for every person in the in the country, but you don't think this can work in the US for a city. You may be quite right that it won't work in the US, but that's a damn shame.