Skip to content Skip to main navigation

Development plans, trams and autonomous mobiles

By Kim Huynh 15 September 2016 57

Autonomous vehicles

Tom Chen and Kim Huynh revisit the best case for light rail and argue that autonomous cars are better

Last week, we wrote an article arguing that light rail is no good for Canberra. Much of the constructive criticism that we received revolved around two points: firstly, that we did not sufficiently take into account the benefits of the projected high density development relating to the tram; and secondly, that we did not provide a detailed alternative.

This article responds to both criticisms.

Light rail is not the only or necessarily best way to improve land value. Indeed, proximity to rail stations has a variable and often meagre impact on property prices. Here’s some of the related studies on the matter.

  • This 2015 article finds that an increased number of bus transit stops also correlated with higher property prices.
  • This study of Brisbane’s transport system found that proximity to both light rail and bus rapid transit had uplift benefits, but only if there were frequent services.
  • This 2007 study and this 2013 analysis of 23 research projects found that the uplift impact of additional light rail services is low if an area is already well serviced by bikes, buses or cars.
  • The most comprehensive 2016 analysis of 140 studies from 60 examples stretching across 40 years in the US concludes that proximity to rapid transit had very mixed outcomes in terms of both increases in land value and the quality of development.

That is why we should be worried that 60% of the government’s reported benefits of light rail are from wider economic and land use benefits, especially when the benefits are  marginal compared to the costs.

We should also be wary that the government is not double counting benefits when it comes to value uplift. Incorporating such benefits when assessing proposed infrastructure is widely regarded as poor analytical practice.

But let’s suppose that house prices go up because of light rail. Who would gain? Experience from other projects gives us reason to believe that the land sales, rezoning laws, and government expenditure on urban renewal associated with light rail will provide lifestyle benefits to a small minority of Canberrans along with a windfall for property developers. All of this will be largely paid for by taxpayers across Canberra.

It would be better for us to invest in driverless vehicles (see infographic below) which offer all Canberrans an effective, exciting and environmentally sound way to get from A to B. The technology is ready. And they could be operating in a similar time-frame to the tram network at a fraction of the cost because the infrastructure is largely in place.

Driverless vehicles would give Canberrans what we want now while also future-proofing us into the twenty-first century.

Figures from the ABS indicate that 88% of people in ACT used private motor vehicles for commuting in 2012. Unpacking the motivations of these people, 7.5% had no access to public transportation at all, 10.7% had to carry equipment or passengers, 19.7% needed a vehicle to before/during/after hours, 22.0% said that services were not available at the right time, 30.5% found public transport too slow and 45.9% preferred the convenience of driving.

But acknowledging that Canberrans use, need and like cars does not mean that we cannot be green or forward-looking when it comes to transportation.

We envision embracing electric driverless vehicles capable of ferrying passengers around without human input. The benefits of driverless vehicles are many: they significantly reduce vehicle accidents by removing human error; they save you from having to drive so that you can rest, read, eat or type; they reduce requirements for parking and driveways; they can transport the disabled and elderly; they improve air quality; you don’t need to pay for a driver; and they take you directly to where you want to go.

Most importantly, driverless vehicles can be networked together into a self-driving taxi fleet that provides the transportation capabilities of private cars without ever having to buy a vehicle or drive yourself.

Some readers might find the prospect of driverless vehicles frightening. But that’s what people thought about elevators and automobiles when they emerged.

This is not science fiction. By the end of 2016, Uber will have 100 self-driving vehicles in Pittsburgh ferrying passengers around. nuTonomy started offering driverless taxis in Singapore as of August. Google’s fleet of self-driving cars have already clocked up 2.4 million kilometres. These are driverless cars that are engaged in real life driving through real traffic. There are 33 major companies committed to rolling out driverless vehicles, many of which have committed to being on the road by 2020.

Shane Rattenbury is thus wrong to assert that driverless cars are at least 15 years away. The “thorny issues with the technology” are being solved, it’s now up to politicians and planners to fashion the appropriate laws and regulations.

A more detailed feasibility of driverless vehicles should be conducted by the government given both its promise and the progress of other cities. We should aim to have a trial of driverless vehicles in Canberra within the next five years. If that works out, we could have a working fleet by 2025.

The value of a fleet of electric self-driving vehicles beats just about any other transport alternative. The $963 million cost of light rail could purchase a fleet of more than 19,260 driverless vehicles at $50,000 each which modelling suggests could provide Canberra with 600,000+ journeys each workday with no passenger having to wait for more than five minutes to get a ride. The cost of recharging an electric vehicle is about 70% cheaper than running a conventional car. We estimate that the electricity cost of providing the entirety of the ACTs passenger travel 2014 though electric vehicles would be less than the annual cost of running ACTION.

There are no doubt people who believe in light rail not so much because of the numbers or evidence, but because they have a vision of a more sustainable and sophisticated Canberra. Autonomous cars are a better way of achieving that vision.

What do you make of the government’s push to increase population density in Canberra generally and along the Northbourne Avenue corridor in particular? Under what if any circumstances would you ride in a driverless car?

Tom Chen works as a research officer at the Australian National University and believes that people are capable of making the right choices when presented with the right information. Kim Huynh is a RiotACT columnist and is also running as an independent for Ginninderra in the ACT election. Check out more on Facebook at gokimbo or GoKimbo.com.au

 


What’s Your opinion?


Please login to post your comments, or connect with
57 Responses to
Development plans, trams and autonomous mobiles
Filter
Showing only Website comments
Order
Newest to Oldest
Oldest to Newest
dungfungus 12:26 pm 19 Sep 16

bringontheevidence said :

AVs will only work without significant congestion problems if the majority of people share…

But my question is, what incentive is there for me to share a vehicle with someone else if I’ll have my own AV already in the garage? One that I can store my stuff in, play my own music, eat if I want, has baby seats for the kids?

What’s the difference between your car and an AV?

Just like the difference between busses and trams in Canberra operating point to point which is none really, but one is cool, the other passe.

bringontheevidence 10:45 am 19 Sep 16

AVs will only work without significant congestion problems if the majority of people share…

But my question is, what incentive is there for me to share a vehicle with someone else if I’ll have my own AV already in the garage? One that I can store my stuff in, play my own music, eat if I want, has baby seats for the kids?

rommeldog56 8:58 am 19 Sep 16

dungfungus said :

How can you possibly make a conclusion that AVs are going to suit your travel needs as they are they are not available yet and no one has any idea how much they will cost.

Sounds like a comment about the Tram stages 1 and 2.

dungfungus 8:25 am 19 Sep 16

Bonkers said :

As a disclaimer, I havn’t read all the posts, so may be repeating items already covered. But I did live in Syndey for a time, and from a work perspective used to love the fact that once on PT, I could fire up the laptop and get stuff done during the journey.

Autonomous vehicles would give me the same benefit. Light rail/trams (whatever you want to call them) doesn’t, as I’m not in the target zone,

I work as a consultant, and as such I need to be at several different sites within a given day. I also have children, so need to be available for school drop off and pickup. The current ACT bus network doesn’t support this. Light rail definitely won’t. Autonomous cars would though.

How can you possibly make a conclusion that AVs are going to suit your travel needs as they are they are not available yet and no one has any idea how much they will cost.

Twelve months ago, promoters of ride sharing were saying their concept would solve all these perceived needs but little has changed.

Your travel needs seem to be the same of about 90% of other Canberrans who use private motor vehicles. Get used to it because it is not going to change.

Innovation 2:59 pm 18 Sep 16

wildturkeycanoe said :

Innovation said :

As for someone’s gripe about following slow vehicles in suburbs. Really! What would this delay cost you in time? 1 minute? 2 minutes?

That minute slower to the first set of traffic lights then adds another cycle of those lights to go through, probably another minute at least. Then you get stopped at the next set, and then the next and so on. Meanwhile, the cars that got through the first set of green lights has scored greens all the way across town thanks to synchronisation. That initial half a minute slower to get that first set of green lights has now added another ten minutes by the time you get to your destination. I have done this often when we take two cars to ferry kids to sports and such. I can arrive five to ten minutes earlier because I got through the first set of green lights.
Now you want to introduce a whole fleet of slow vehicles to really make it hard for everyone to get where they want to go. Honestly, what is the point of having different speed limits at all? Why don’t we just go back to the days of horse and carriage and do no more than 40km/h everywhere, regardless of the type of road? At least we’ll have fewer casualties, but people will miss out on so much of their free time, spending it on traveling very slowly everywhere.
What is the point of making dual lane highways, then putting a limit on the speed you can do? They were built to get people and goods to their destination quicker, but then we put a restriction on how quick.

Um, I think you’ve misread my posts. I was suggesting A/V’s only to ferry passengers within suburbs to and from nearby main roads. I can’t think of many suburban roads with traffic lights and I doubt there would be any that are synchronised with other traffic lights.

I would have thought that any improvement on existing public transport and which encouraged road users to pull over at the first opportunity and utilise (eg by catching more cost effective multi occupant A/Vs or, for the main roads, a Direct/Rapid driver operated ACTION bus would reduce congestion. In turn, overall travel times for those people who still need to drive (or prefer to pay more to drive) might be less than they are now.

wildturkeycanoe 12:17 pm 18 Sep 16

Innovation said :

As for someone’s gripe about following slow vehicles in suburbs. Really! What would this delay cost you in time? 1 minute? 2 minutes?

That minute slower to the first set of traffic lights then adds another cycle of those lights to go through, probably another minute at least. Then you get stopped at the next set, and then the next and so on. Meanwhile, the cars that got through the first set of green lights has scored greens all the way across town thanks to synchronisation. That initial half a minute slower to get that first set of green lights has now added another ten minutes by the time you get to your destination. I have done this often when we take two cars to ferry kids to sports and such. I can arrive five to ten minutes earlier because I got through the first set of green lights.
Now you want to introduce a whole fleet of slow vehicles to really make it hard for everyone to get where they want to go. Honestly, what is the point of having different speed limits at all? Why don’t we just go back to the days of horse and carriage and do no more than 40km/h everywhere, regardless of the type of road? At least we’ll have fewer casualties, but people will miss out on so much of their free time, spending it on traveling very slowly everywhere.
What is the point of making dual lane highways, then putting a limit on the speed you can do? They were built to get people and goods to their destination quicker, but then we put a restriction on how quick.

dungfungus 11:32 am 18 Sep 16

I was just watching on TV some of the AVs being trialled in the USA – hilarious.

With all that spinning stuff and satellite arrays on top of the vehicle they would be useless for going to Bunnings to get wood etc. because there is no room for a roof-rack.

I suspect the boot space is full of electronic stuff also (like an oversized battery to power everything). Then there will be the batteries for the hybrid propulsion.

Very un-Australian, I think.

Bonkers 12:00 am 18 Sep 16

As a disclaimer, I havn’t read all the posts, so may be repeating items already covered. But I did live in Syndey for a time, and from a work perspective used to love the fact that once on PT, I could fire up the laptop and get stuff done during the journey.

Autonomous vehicles would give me the same benefit. Light rail/trams (whatever you want to call them) doesn’t, as I’m not in the target zone,

I work as a consultant, and as such I need to be at several different sites within a given day. I also have children, so need to be available for school drop off and pickup. The current ACT bus network doesn’t support this. Light rail definitely won’t. Autonomous cars would though.

Innovation 10:30 am 17 Sep 16

To the naysayers of my earlier post, my suggestions were merely an attempt to find a middle ground where Canberra could start trialling AVs.

Initial slow (eg 20km/h) limits would minimise the risk of accidents and injury – and the degree of damage if an accident occurred.

The cost of small AVs in suburban areas only – ie ferrying passengers to and from main driver operated transit corridors (aka “the last mile” as mentioned by someone else on the his thread) – would be significantly less than the cost of our current oversized driver operated diesel buses, so AV frequency could be increased. Passengers would not have to wait or even plan trips potentially making their overall trip faster.

If A/V and driver operated bus trips became cheaper than multi occupant car trips (or even free), more car drivers and/or car passengers might start pulling over at the first opportunity and switching to buses for part or all of their trip.

As for someone’s gripe about following slow vehicles in suburbs. Really! What would this delay cost you in time? 1 minute? 2 minutes? No doubt you waste more time watching commercials on TV each day or driving around looking for car parks.

Oh and having an app that broadcasts a type of vehicle’s proximity (and could be automatically linked to the ignition or sirens etc) would be very different to a device that broadcasts traffic delays which would require more significant, and geographically widescale electronics coupled with human intervention and operation.

And the legal issues could be resolved easily. AV’s are riddled with cameras. I’m not aware of there ever having been doubt as to who was at fault with an AV. I’m sure the AV designers – keen to get their vehicles on the road – would have accept a contractual agreement with the ACT Government as to who would assume liability if damage or injury was caused by an AV.

And on top of all this, the ACT Government is in the perfect position of having sole jurisdiction to implement any road rule changes, sign contracts, reorganise public transport and manage insurance. As I mentioned before, the cost of allowing a few companies to trial their AVs in select suburbs might even be negligible just so those companies could get a foot in the door.

Even bus drivers and unions should be somewhat happy with this approach. Existing buses and drivers could be repurposed to operate added or more frequent main road/direct bus routes (eg, as per the Liberals’ policy). They would transport many more passengers than the sometimes negligible (or even nil) numbers in suburbs now and they would move passengers much more quickly than they do now. (They wouldn’t even have to put up with impatient drivers tailgating them and complaining about how slow they are 😉 ). Surely that would increase job satisfaction.

KentFitch 9:51 am 17 Sep 16

K320Scania said :

You can go on about how great Canberra’s road system is, but that’s the problem. Canberra’s reliance on cars and roads is just ridiculous. What angers me most about those who say Canberra has an excellent road system, is that they don’t understand what a huge impact these roads have had on the local environment – it’s sad to hear these people say, “there’s a dead kangaroo on the side of the road,” – it’s become an expectation to see dead wildlife on our roads, as if they’re worthless creatures that we don’t have to care about. Roads devide an area of land making kangaroos get bordered in to a small pocket, and then when they try to cross to go back to their former area, now divided by a road and cross it, they get hit by a motor vehicle, because animals simply can’t judge high speeds.
Canberra may have an “excellent” road system, but it has come at a cost of so much of our environment – killed one hundred year old trees, killed kangaroos, wombats and other amazing animals that you just don’t get to see elsewhere.
Instead of the bush capital as Canberra’s nickname, it should be changed to the bush destruction capital, to reflect what we’ve done to it.
Autonomous cars will demand for more lanes of road, new roads and new road infrastructure, at the cost of our environment, and that basically means more environment abuse, which I find simply unacceptable. If you can’t respect Canberra’s unique environment (bush at your doorstep), then go live in Sydney or Melbourne – cities dominated by cars and congestion, with no trees or greenery and controlled by the motor car and its needs – eyesore eight lane roads, which came at a cost of the environment.

One of the main reasons I am an advocate for autonomous cars is because they will reduce the environmental and economic costs of transport. Autonomous cars allow road infrastructure to be used much more efficiently – we will need fewer road lanes and be able to release land/space used by car-parks and garages for more “productive” uses (parks, pseudo-bush, shops, housing? – it should be a community decision).

I know Parsons Brinckerhoff are a multi-national engineering consultancy and what they say has to be interpreted at least partly self- promotion, but it is worth remembering that until recently, they were seen as part of the heavy-engineering-heavy-transport-infrastructure lobby: big roads and rail lines (Lane Cove tunnel, Ipswich Motorway, …) Now see what they are promoting – their “Making better places” whitepaper is filled with before-and after imagined representations of city and urban and arterial roads showing how autonomous vehicles can return land from roads [ http://www.wsp-pb.com/Globaln/UK/WSPPB-Farrells-AV-whitepaper.pdf ] This document is targeted at UK authorities. They estimate 8,000 hectares of central London land is used for car parking, noting that a single 100 hectare redevelopment on this land is worth about 1.25 billion pounds in “value uplift”. They also note that car parking continues to encroach on residential gardens: “in 2015 five times as many London homes had front gardens with no plants compared with 2005” (I found it surprising that they note 44% of inner London households do not have a car, because that implies 56% of them do have a car!)

The economic and environmental benefits in Canberra of not needing to continually widen roads and to be able to release car-parks to better uses are significant and long-term. What would that multi-story-monstrosity hiding the Belconnen Mall from the lake become? After a few years, would you convert your unneeded garage into a rumpus room, workshop, home theatre or an emergency housing shelter? Would you plant tomatoes on your old driveway?

dungfungus 9:45 am 17 Sep 16

Here is a link to a report of testing that is happening in Pittsburgh right now.
I think Uber has a long way to go . I can’t see AVs working with anything else but other AV’s.

https://nextcity.org/daily/entry/uber-tests-self-driving-cars-pittsburgh-drivers

MarkE 11:33 pm 16 Sep 16

I strongly agree with these conclusions about driverless cars being a better solution than light rail. Light rail is a 19th century solution to a 21st century problem.

“Improving land values” is not something the Government should involve itself in other than to remind people that Canberra has a leasehold property system and land was supposed be handed back to the Government after 99 years to prevent land speculation in Canberra.

Why should a baby boomer get to renew a 99 year lease in Forrest for a nominal sum when a young person pays hundreds of thousands of dollars for a lease on a tiny block on the fringes of Canberra? Doesn’t seam fair.

In the 1970’s land for housing sold in Belconnen for as little as $1.

As for the massively over-priced property values in Australia. There has never been debt fueled asset bubble that did not crash. Interest rates are at their lowest levels in 5,000 years. The super low interest rate thing is just an experiment by Reserve Banks that has already produced bubbles and busts in shares, housing and the biggest bubble in history, the bond market. The problem with over priced property WILL self correct.

Mark Ellis is a Liberal Democrats Candidate for the seat of Kurrajong

K320Scania 8:22 pm 16 Sep 16

You can go on about how great Canberra’s road system is, but that’s the problem. Canberra’s reliance on cars and roads is just ridiculous. What angers me most about those who say Canberra has an excellent road system, is that they don’t understand what a huge impact these roads have had on the local environment – it’s sad to hear these people say, “there’s a dead kangaroo on the side of the road,” – it’s become an expectation to see dead wildlife on our roads, as if they’re worthless creatures that we don’t have to care about. Roads devide an area of land making kangaroos get bordered in to a small pocket, and then when they try to cross to go back to their former area, now divided by a road and cross it, they get hit by a motor vehicle, because animals simply can’t judge high speeds.
Canberra may have an “excellent” road system, but it has come at a cost of so much of our environment – killed one hundred year old trees, killed kangaroos, wombats and other amazing animals that you just don’t get to see elsewhere.
Instead of the bush capital as Canberra’s nickname, it should be changed to the bush destruction capital, to reflect what we’ve done to it.
Autonomous cars will demand for more lanes of road, new roads and new road infrastructure, at the cost of our environment, and that basically means more environment abuse, which I find simply unacceptable. If you can’t respect Canberra’s unique environment (bush at your doorstep), then go live in Sydney or Melbourne – cities dominated by cars and congestion, with no trees or greenery and controlled by the motor car and its needs – eyesore eight lane roads, which came at a cost of the environment.

OpenYourMind 6:31 pm 16 Sep 16

wildturkeycanoe said :

Innovation said :

Canberra would be an ideal place to trial autonomous buses!

Small, approx ten seater electric vehicles, could be trialled by replacing a handful of current slow and circuitous suburban bus routes to ferry passengers to main transport corridors. I think suburban buses average around 25km/h so the replacement buses, which would be more frequent, could be speed limited to around 20km/h and still provide a better service. (I’m sure that someone could create an app even that alerted people to the proximity of these vehicles – as already should be the case for emergency service vehicles.)

OpenYourMind said :

where the Google car takes a very conservative and safe approach and human drivers don’t anticipate that!

So basically we are advocating putting more ultra slow drivers [in this case instead of silver hair, they have silver outer shells] onto our roads. These vehicles will wander around the suburbs, doing well under the speed limit, stopping frequently and basically making a right nuisance of themselves to all other road users. Instead of me being able to take a 20 minute journey into civic in my own fast car, I will be either a passenger to one of these mini buses that meander around the burbs picking up other souls who have to get ready an extra half an hour early to get to work on time, or I am stuck behind one, cursing at the inability to overtake anywhere because of the double whites along the entire route.
Buses already do this now, putting along doing 25km/h because they are ahead of schedule due to not enough passengers. You can’t overtake them on back streets whilst they are moving due to oncoming traffic or double lines, nor can you go around when they are stopped because there is usually a traffic island right next to it. Have Canberra planners done this deliberately to frustrate motorists?

Nope, you’ve missed the subtlety. The driverless cars aren’t driving particularly slowly, they are simply obeying the speed limits, indicating, stopping at stop signs and generally driving in a cautious and safe manner. This is what is catching other drivers a little off guard! A driverless car could consistently outlap you on a racetrack if you want one to travel as fast as possible. Some of the experimental cars such as the Google bubble car may be slower during developmental phases, but overall, the expectation is that the average speed of driverless vehicles will be quite a bit faster than humans. Reason being that most delays in traffic are human behaviour related. A network of driverless vehicles that communicate with each other and intelligently route around problem areas will perform better than the same number of human drivers.

dungfungus 11:02 am 16 Sep 16

KentFitch said :

wildturkeycanoe said :

So basically we are advocating putting more ultra slow drivers [in this case instead of silver hair, they have silver outer shells] onto our roads. These vehicles will wander around the suburbs, doing well under the speed limit, stopping frequently and basically making a right nuisance of themselves to all other road users. Instead of me being able to take a 20 minute journey into civic in my own fast car, I will be either a passenger to one of these mini buses that meander around the burbs picking up other souls who have to get ready an extra half an hour early to get to work on time, or I am stuck behind one, cursing at the inability to overtake anywhere because of the double whites along the entire route.
Buses already do this now, putting along doing 25km/h because they are ahead of schedule due to not enough passengers. You can’t overtake them on back streets whilst they are moving due to oncoming traffic or double lines, nor can you go around when they are stopped because there is usually a traffic island right next to it. Have Canberra planners done this deliberately to frustrate motorists? This isn’t made up either, I have many times had my journey to the local shops slowed down to the pace of the local bus, obviously running early as there was indeed no attempt to speed up to anything over 30km/h. They are almost as bad as three cyclists riding abreast.
The “statewide” multi-vehicular approach to infrastructure is not going to improve people’s journeys. Presently you walk to the local stop, get the bus to the nearest interchange, jump onto the next express, to the next interchange, to the next local service and then you walk the last few hundred meters or kilometers to your destination. Just because you are taking a mini bus or autonomous vehicle for the first part of the journey, it won’t eliminate the need for the rest of the network or make it any faster. Booking an autonomous vehicle from your phone [because they won’t be driving around on pre-specified routes like buses], will mean you need to wait for a car that is doing a similar journey to yours. It won’t be parked near your house so you have to wait for it to come from wherever they are stored during the day/night. It has to contend with the same slow buses I just mentioned, plus it is also not going to drive aggressively to get to your home faster. There is many wasted minutes before you even leave. Then the same over-cautious car will drive you to the bus interchange, deviating around some side streets to pick up more fares, making your previously five minute DIY trip now up to three or four times slower.

I cannot see an improvement in travel times, but rather a degradation of our current transport system, clogged up with paranoid AIs that are too scared to do the speed limit.

“I’m sure that someone could create an app even that alerted people to the proximity of these vehicles – as already should be the case for emergency service vehicles.”
There isn’t even an app that warns of road congestion in peak periods now, let alone tracking of our buses. The NXTbus page had no real time vehicular tracking or movement in the 5 minutes I stared at it. The only thing it could tell me was that it’d take me nearly an hour to get to Civic, which I can do in half that driving my own car.
All this talk of new technology and we can’t get one working that was installed how many years ago?

If a shared fleet of AVs did works as you suspect, it will be an abject failure, and the tens of billions being invested in it will be wasted. But those investing are doing to to build an on-demand service better (including faster, by reducing congestion and synchronising traffic) and cheaper than current private cars and public transport. They may be deluded, in which case they will lose those billions, but you must at least acknowledge that they (Ford, GM, Toyota, BMW, Audi, Volvo, Tesla, Google, Baidu, …) think they are onto something!

What if they are right? How do we calculate the odds they are right? I suggest we look at what the academics and researchers are saying, what the think-tanks and engineering consultancies and bankers are writing, and run own own ruler over the results, build models we can configure and run simulations, and make better informed guesses, that’s all.

I don’t have any ulterior motive – Im not an investor or employee of any of these businesses. My motivation is better mobility for me, my family and friends and exploring whether social equity can be promoted by universal and cheap transport.

You may not accept the assumption behind my model, which I have attempted to make explicit: http://canberraautonomouscars.info/model.html

But if you do, run the default simulation http://canberraautonomouscars.info/sim.html and when it finishes, scroll down to the “Traces of 3 random cars” section to see some typical trips. If you scroll down further, you can see “10% sample of traces of visits to Giralang” (just as an example).

“If a shared fleet of AVs did works as you suspect, it will be an abject failure, and the tens of billions being invested in it will be wasted. “

Exactly the sort of project the current ACT Government would love.

It would be like Rhodium Asset Solutions on steroids.

devils_advocate 11:01 am 16 Sep 16

KentFitch said :

If a shared fleet of AVs did works as you suspect, it will be an abject failure, and the tens of billions being invested in it will be wasted. But those investing are doing to to build an on-demand service better (including faster, by reducing congestion and synchronising traffic) and cheaper than current private cars and public transport. They may be deluded, in which case they will lose those billions, but you must at least acknowledge that they (Ford, GM, Toyota, BMW, Audi, Volvo, Tesla, Google, Baidu, …) think they are onto something!

Well another key difference there is that these investments, by and large, are by private people making their own private investment decisions with their own private money. That’s very different than committing an entire town to pay for something, regardless of whether they want or can afford it, and even worse something that only a very few of them will see the benefits of.

devils_advocate 10:58 am 16 Sep 16

OpenYourMind said :

Almost every car today would be susceptible to an ‘EMP bomb’. Our entire society operates with the vulnerabilities associated with electronics.

Correction: every car built today. Plenty of cars getting around today (including one of mine) that runs off carburettors, throttle cables, hydraulic brakes and clutches, with nary a computer to be seen.

KentFitch 10:25 am 16 Sep 16

wildturkeycanoe said :

So basically we are advocating putting more ultra slow drivers [in this case instead of silver hair, they have silver outer shells] onto our roads. These vehicles will wander around the suburbs, doing well under the speed limit, stopping frequently and basically making a right nuisance of themselves to all other road users. Instead of me being able to take a 20 minute journey into civic in my own fast car, I will be either a passenger to one of these mini buses that meander around the burbs picking up other souls who have to get ready an extra half an hour early to get to work on time, or I am stuck behind one, cursing at the inability to overtake anywhere because of the double whites along the entire route.
Buses already do this now, putting along doing 25km/h because they are ahead of schedule due to not enough passengers. You can’t overtake them on back streets whilst they are moving due to oncoming traffic or double lines, nor can you go around when they are stopped because there is usually a traffic island right next to it. Have Canberra planners done this deliberately to frustrate motorists? This isn’t made up either, I have many times had my journey to the local shops slowed down to the pace of the local bus, obviously running early as there was indeed no attempt to speed up to anything over 30km/h. They are almost as bad as three cyclists riding abreast.
The “statewide” multi-vehicular approach to infrastructure is not going to improve people’s journeys. Presently you walk to the local stop, get the bus to the nearest interchange, jump onto the next express, to the next interchange, to the next local service and then you walk the last few hundred meters or kilometers to your destination. Just because you are taking a mini bus or autonomous vehicle for the first part of the journey, it won’t eliminate the need for the rest of the network or make it any faster. Booking an autonomous vehicle from your phone [because they won’t be driving around on pre-specified routes like buses], will mean you need to wait for a car that is doing a similar journey to yours. It won’t be parked near your house so you have to wait for it to come from wherever they are stored during the day/night. It has to contend with the same slow buses I just mentioned, plus it is also not going to drive aggressively to get to your home faster. There is many wasted minutes before you even leave. Then the same over-cautious car will drive you to the bus interchange, deviating around some side streets to pick up more fares, making your previously five minute DIY trip now up to three or four times slower.

I cannot see an improvement in travel times, but rather a degradation of our current transport system, clogged up with paranoid AIs that are too scared to do the speed limit.

“I’m sure that someone could create an app even that alerted people to the proximity of these vehicles – as already should be the case for emergency service vehicles.”
There isn’t even an app that warns of road congestion in peak periods now, let alone tracking of our buses. The NXTbus page had no real time vehicular tracking or movement in the 5 minutes I stared at it. The only thing it could tell me was that it’d take me nearly an hour to get to Civic, which I can do in half that driving my own car.
All this talk of new technology and we can’t get one working that was installed how many years ago?

If a shared fleet of AVs did works as you suspect, it will be an abject failure, and the tens of billions being invested in it will be wasted. But those investing are doing to to build an on-demand service better (including faster, by reducing congestion and synchronising traffic) and cheaper than current private cars and public transport. They may be deluded, in which case they will lose those billions, but you must at least acknowledge that they (Ford, GM, Toyota, BMW, Audi, Volvo, Tesla, Google, Baidu, …) think they are onto something!

What if they are right? How do we calculate the odds they are right? I suggest we look at what the academics and researchers are saying, what the think-tanks and engineering consultancies and bankers are writing, and run own own ruler over the results, build models we can configure and run simulations, and make better informed guesses, that’s all.

I don’t have any ulterior motive – Im not an investor or employee of any of these businesses. My motivation is better mobility for me, my family and friends and exploring whether social equity can be promoted by universal and cheap transport.

You may not accept the assumption behind my model, which I have attempted to make explicit: http://canberraautonomouscars.info/model.html

But if you do, run the default simulation http://canberraautonomouscars.info/sim.html and when it finishes, scroll down to the “Traces of 3 random cars” section to see some typical trips. If you scroll down further, you can see “10% sample of traces of visits to Giralang” (just as an example).

KentFitch 10:07 am 16 Sep 16

gooterz said :

Kim do you actually know the status of Autonomous vehicles?

Roadwork. You can’t use them
Unmapped roads you can’t use them.
Any sort of change the car doesn’t expect the user has to take over control.

If you want a decent car it’ll probably cost you 60k extra for the system which isn’t released yet and not approved in Australia at all.

Who has public liability if a car hits someone? is ACT government going to pay for every accident?

Perhaps if instead of Autonomous vehicles we employed drivers for the cars. Made the cars bigger and sold tickets. Then instead of random routes just fixed some routes in place that people would normally be able to get to. You could call it a bus.

Have you seen that Action ridership is down again this year?

I do not believe AVs will be able to meet every possible transport needs for decades. Rather, in 5 years, it is very likely that they will work very very well in constrained circumstances, such as urban Canberra and travel on main trunks.

Coping with road works, both planned and emergency, are a known design requirement: they will be noticed by the first AV to encounter them, if not already added to the shared route information database, and each AV is updated with the new information.

Coping with unmapped roads (not on GoogleMaps for example?) is not a design requirement for providing the vast bulk of transport in urban Canberra (in the spirit of not letting the perfect be the enemy of the good).

Ford are planning for shared fleet AV without steering wheel/pedals in 2021 – you wont be able to “take control” other than specify a destination or invoke an emergency stop.

Will they cost 60K? Almost certainly not, as the electronics and sensors, as with all mass produced digital equipment, will be very cheap.

Insurance and liability is a great question that we should be discussing. Volvo assert they will be liable, but maybe they are talking big.

Yes, ACTION ridership as a percentage of trips continues to slide, even with pay parking hikes. Why is that? I assert it is because people highly value door-to-door, on-demand, 24×7 mobility, and ACTION (with or without trams) does not provide this. When I lived in EmuRidge, I was lucky that taking an ACTION bus was very convenient for most trips, but now, it adds over 2 hours per day to my typical commute, and the tram will actually make it worse by adding extra transfers (I now live in Nicholls and work sometimes at UNSWCanberra/ADFA and sometimes in Parkes.)

KentFitch 9:50 am 16 Sep 16

bringontheevidence said :

Kim and Tom, I’d like to address a few technical issues around the economics of light rail, benefit/cost analysis and the economics of ‘alternatives’ of light rail that you may have neglected. Sorry for the excessive detail.

1. Driverless vehicles are not an alternative to peak period mass transit. In fact, the benefits of driverless vehicles (lower effort car transport) are likely to make mass transit more necessary because of increased peak demand for roads.
Also, you talk about the ‘costs’ of driverless vehicles being low because a fleet of 20,000 could offer 600,000 journeys per day. While this might be technically correct, that would mean nearly full utilisation of the whole fleet 24 hours a day and assumes the vehicles are located close to where they need to be at all times. In reality, demand would significantly peak in the mornings and afternoons, requiring a significantly larger fleet of vehicles to meet demand. These vehicles would then sit idle 20+ hours per day.
Further, driverless cars still take up road capacity so to cater for all of those extra trips you would need a massive road investment program which would cost almost as much as the cars themselves. Especially given once you reach a certain saturation of roads in a city, additional road capacity will only come at great cost.

As a result, good quality and high capacity public transport will still be required to service peak demand. To this end Canberra DOES need either light rail or BRT. So the question is, which is better?

Your intuition is similar to mine after I read Columbia University’s Earth Institute report, “Transforming personal mobility” [ http://sustainablemobility.ei.columbia.edu/files/2012/12/Transforming-Personal-Mobility-Jan-27-20132.pdf ] in which a simulation of a shared fleet of AVs in Ann Arbor showed a very small number of vehicles could service the transport needs serviced by private cars and public transport will a door-to-door, on-demand service. It didnt make intuitive sense. And their traffic flows were much more symmetrical than Canberra’s “tidal” commuter flows: towards Civic/Parkes in the morning and back to the suburbs at night.

A spreadsheet simulation suggested I might be wrong – that a small fleet would work as long as cars were shared during peak periods, so there was only one thing to do, build a realistic traffic model using known traffic patterns and volumes and an economic model to see what the break-even cost would be. The model and its assumptions are documented here: http://canberraautonomouscars.info/model.html and you can change most of them when running the simulation. For example, the model assumes AV cost $40K, are financed with money costing 10%, have a 3 year life and are then written off ($0 scrap value). If you think they’ll cost $50K, finance costs 7% and have a $5K scrap value, plug that in (and you’ll see the operating surplus is not much different at the 750K journeys/day simulation). But I assert the default settings are conservative. Independent models from UTexas (eg http://www.ce.utexas.edu/prof/kockelman/public_html/TRB14SAVenergy_emissions.pdf http://www.caee.utexas.edu/prof/kockelman/public_html/TRB15SAVsinAustin.pdf ) and Berkeley Labs ( http://newscenter.lbl.gov/2015/07/06/autonomous-taxis-would-deliver-significant-environmental-and-economic-benefits/ ) show similar or more “optimistic” results.

Note that for 23K cars to service 750K daily journeys, they must be shared in peak periods: in peak periods, you hire a seat, and on highly trafficed routes (eg, Kambah to Parkes) you will almost certainly be sharing the car with 1, 2 or 3 other people, also travelling from near Kambah to near Parkes (with the model parameter that sharing will never make your trip more than 5 minutes longer than it would otherwise take). Out of weekday peak (and all weekend) you hire the whole car (and at a cheaper rate with the default model).

I agree it is not intuitive that 23K cars can service all public transport and normal car needs in Canberra (excluding tradies’ utes, horse-floats etc). If you look at the output of a typical 750K journey simulation, you’ll see that even at highest peak times, about 3000 cars are idle (not in use, not be relocated to meet anticipated demand, but empty, waiting for something to do). Distributed around about 110 “suburbs/localities”, that’s an average of about 25 per suburb, waiting for a request. And each minute, over a thousand vehicles complete a journey or relocation.

Again, for a typical 750K journey simulation, each car is idle for about 43% of the day. Private cars are idle for about 97% of the day. That is, even if you could double the utilisation of the private car, it would be idle for 94% of the day; increase the utilisation of the private car by ten times, and it would be idle for 67% of the day. Unintuitive, I agree. As well, the average private car in Canberra has an occupancy of just 1.2 people for that busy 3%

In comparison, the average AV arriving in Civic/Parkes/Barton in the AM peak is carrying 2 people. As you can calculate, this takes a huge number of vehicles off the road, and in Canberra’s case, eliminates congestion. As demand grows, the number of cars needed (the fleet size) grows much more slowly, as average occupancy grows.

bringontheevidence said :

2. The Government’s benefit cost analysis found that the benefit of BRT was similar to light rail, but cost significantly lower, so it had a better benefit/cost ratio. This outcome ignores a couple of methodological issues with the benefit/cost analysis and how it’s been used in this case.
Firstly, the comparison assumed the same passenger numbers for both light rail and BRT. This was based on the assumption that because travel times and cost would be the same, the passenger numbers would also be the same. However, evidence from overseas shows that even when travel times and cost are the same, passengers show a distinct ‘rail preference’, indicating passenger numbers for light rail would be larger than for BRT (improving the relative benefits of light rail, particularly in primary transport benefits).
Secondly, the Australian standard for benefit/cost analysis requires analysts to use a discount rate much higher than elsewhere, whereas in the UK, for example, the discount rate is much lower . Using such a high rate means that future benefits are valued much lower in Australia, while current costs are much higher (relatively). This means two things in practice. Firstly, infrastructure investment in Australia is undervalued in the benefit/cost model because the discount rate doesn’t reflect the cost of capital to the Government (which is at historic lows). Secondly, the high rate also means benefit/cost ratios are biased towards lower capital cost/high operating cost options (like BRT) vs high capital cost/lower operating cost options (like light rail).
If these methodological issues were addressed, light rail would stack up much, much better compared to BRT in the current model, and the overall benefit/cost of the project would be significantly higher (the Gov ran the project using the UK standard BCR model and the benefits outweighed the costs 2-1.

The Capital Metro business case actually factored-in very generous travel time benefits which were not supported by the their EIS. Specifically, their EIS modelled increased travel times in the corridor containing the tram, which should have been a negative travel time benefit in the business case. That is, the EIS showed NEGATIVE travel time savings, and hence magic multipliers and “uplifts” based on better travel outcomes in the business case are fanciful. [ http://canberraautonomouscars.info/faq.html#time1 http://canberraautonomouscars.info/faq.html#eisapend ]

How can this be? The main reason is that the tram route crosses 23 intersections. Many of these intersections are very busy with lots of cross-traffic (which includes buses). If the tram is given signal priority, signal synchronisation along the route breaks down for traffic flowing along the route and cross traffic is greatly slowed. Here’s what happens at Mouat St using the Capital Metro EIS signal timings and traffic volumes, now and with and without the tram in 2021 and 2031 : http://canberraautonomouscars.info/simMouat.html

CBR Tweets

Sign up to our newsletter

Top
Copyright © 2019 Region Group Pty Ltd. All rights reserved.
the-riotact.com | aboutregional.com.au | b2bmagazine.com.au | thisiscanberra.com

Search across the site