Warwick Cathro [1]

warwickcathro@gmail.com

Farrer, Australian Capital Territory

INTRODUCTION

In a keynote presentation [2] at the 2023 AEVA National Conference, Brent Toderian (an eminent Canadian urban planner) told us that our cities and suburbs “need fewer cars, smaller cars and less driving”.

His presentation was a plea for our urban areas to be re-designed as “compactable, walkable and complete communities” where most journeys are made by active travel (principally walking and cycling) or by public transport, and not as communities “where cars are regarded as the only viable way to get around”.  

For decades AEVA has been advocating the replacement of fossil fuel vehicles with electric vehicles. But many argue that one-for-one replacement is not the best outcome. Would it be a better outcome if we had “half the cars, but all of them electric” [3]?

Advocates for sustainable transport [4] strongly argue for a shift away from private car use and towards greater use of public transport and active travel. For example, Climate Action Network Australia (CANA) stated in 2024 [5]: 

“Enabling mode shift … away from the use of private cars is essential to reduce emissions … and also would improve health outcomes, reduce congestion, save people money, create more jobs, and contribute to more connected communities”.

OBSERVATIONS

While I agree with the aspiration for fewer cars, this transition encompasses a number of issues which need to be unpacked. I will comment on seven of them. Some of these comments relate particularly to Canberra where I live, a city with a relatively low population density, but many of them will apply to other cities. 

Land use

Brent Toderian drew our attention to the climate benefits, the health benefits and the land use benefits of fewer cars. The “land use benefits” are the productive outcomes that communities could realise by replacing urban freeways and parking expanses.

Australians, through expenditure by all levels of government and through tolls, have been spending more than $30 billion per year on roads, including urban freeways and wider roads [6]. Dual carriageways and freeways are being constructed in our cities to cope with growing peak hour congestion due to a high level of car use for commuting.  

Photo credit: [P1]

There is also the space consumed by parking. In her seminal 1961 book The death and life of great American cities [7], in addition to her advocacy of “dense and diverse neighbourhoods”, Jane Jacobs commented on the “duplication of car parking”: the schools, the supermarkets, the churches, the shopping centres, the clinics, the movie, all the residences, must have their own parking lots and all this duplicate parking lies idle for much of the time.   

Commuter travel

A significant component of the land use challenge highlighted by Brent Toderian relates to commuter travel. According to the 2016 Census, driving is the dominant method of travel to work in Australia, with 69% of the working population commuting by car as a driver, and a further 5% as a passenger. Roads, like the electricity grid, are being designed to handle peak load. If privately owned cars commuting to and from work carried an average of (say) 2.2 people, instead of 1.1 people, the number of cars on our roads during peak hours would be halved. Such an outcome is both highly desirable and difficult to achieve.

Car idle time

Fewer cars need not mean fewer car journeys, or fewer kilometres travelled. It may mean that there are fewer, but more intensively used cars. Studies have shown that our cars are idle up to 95% of the time [8]: they are parked at home, at work or at a carpark near another destination. Brent Toderian advocates “driving less” but his land use goals, particularly those related to parking, would arguably be met if the same number of journeys could be made with fewer cars.

Definition of public transport

The advocates of a shift away from car use are not always clear on what they mean by “public transport”. Does this term include on-demand services, such as AT Local in Auckland [9]?  Does it include taxi and Uber services?  Some definitions of “public transport” include a requirement for “fixed routes”, but the definition does not matter as long as the advocates of mode shift are including these on-demand options in their advocacy. 

Convenience

Those who advocate for this mode shift do not always set out a mechanism for how it can be achieved, and rarely discuss the variety of journey types that we all undertake, some of which are far more conveniently undertaken by car.

A journey to the local shops for a few grocery items certainly can and should be made on foot or by bicycle, by all those for whom disability or distance is not an obstacle. But doing a weekly grocery shop for a family is a different matter, unless it becomes a universal practice to grocery shop online and have our groceries delivered [10]. 

Parents frequently use cars to transport their children to a variety of places and events, such as school, childcare and weekend sporting commitments. Where these events involve multiple children and different suburbs across town, public transport and active travel can be markedly less convenient than the family car. 

My wife and my step daughter have raised young children in Canberra, a city with a relatively low population density. I asked them if they thought it would be possible to be a working parent of young children in Canberra and not own a car. They both said “it would either be impossible or extremely stressful”.  

Park-and-rides

Commute journeys are not single-mode journeys. Many of them involve multiple modes such as walking, car travel and fixed-route public transport. While advocates of sustainable transport are not fond of car parks, the fact is that park-and-ride car parks can play an important role in encouraging a shift to fixed-route public transport, especially in outer suburbs which lack bus feeder routes. If park-and-rides were to instal large numbers of ordinary powerpoints for EV drivers, we might see three benefits: (a) an encouragement to make public transport a large component of commuter travel; (b) a solution for EV drivers who are unable to charge at home; and (c) the consumption of electricity when it is most needed, during the solar-created surplus in daylight hours.

Expense

Car ownership is expensive.  Even people who only ever buy used cars face significant costs in registration and insurance. Some people eschew car ownership and rely on bicycles or electric scooters in combination with public transport, creating an imperative to upgrade active travel infrastructure. Other people may rely largely on Ubers or car-sharing services to get around. This brings us to the issue of car sharing.

CAR SHARING AND CAR POOLING

There are two main types of car sharing services in Australia: car subscription services such as GoGet, Carly or Carbar, where the service owns the cars, and owner-supported services such as Drive Mate and Turo, where individual owners rent out their cars.

The subscription services are based on a weekly or monthly car subscription that drivers can change at short notice. GoGet has been one of the most enduring of the subscription services, but it is difficult to find recent data on its membership [11]. 

A few companies have launched owner-supported carshare services. Uber launched such a service in 2022, but closed it in 2024 [12]. Turo and Drive Mate continue to operate.

Under this model, car owners can offer their car after submitting data about their vehicle and undertaking an orientation interview. Importantly, comprehensive insurance cover is provided for all trips made through these services. While the traditional Uber taxi-like service has been a success, it remains to be seen whether Australian drivers will choose to make money from their cars by participating in owner-host carshare services.

The subscription model can support people who have decided not to own a car and who routinely cycle or use public transport, but require a car for the occasional long journey or to transport heavy goods. 

Voluntary car-pooling for commuters is impeded by the fact that employees currently work with highly varying schedules. As one commenter on a Reddit discussion page [13] said: “the odds of finding someone who lives close to you and works the same schedule as you are pretty slim”. People are reluctant to participate in an arrangement that might make them late for work, require them to leave work at an inconvenient time, or make it impossible to do grocery shopping on the way home. 

DYNAMIC CARPOOLING

Traditional carpooling requires that journeys are planned well in advance. In theory, a promising alternative is dynamic carpooling (sometimes called “instant carpooling”) where commuters find travel partners spontaneously, often at the last minute, using mobile applications or online platforms [14]. The people you pool with are different for every journey, and are selected based on the times and on the destinations that they happen to be travelling on a particular day.

Uber advertises a dynamic carpooling service called Uber Commute [15]. It promotes this service as “reducing traffic” and “reducing the cost of your commute”. The service matches pairs of neighbours and co-workers travelling similar routes. These commuters must submit their morning commute request by 9pm the night before, and their evening commute request by 3pm the same day.

It is difficult to discover much public information about the use of Uber Commute in Australia. It seems to have attracted very little notice.

ROBOTAXIS

In planning future solutions for sustainable transport, local governments could examine the potential connection between dynamic carpooling and robotaxis. 

Robotaxis, or “driverless taxis”, are applications of autonomous vehicle technology, in which fleets of driverless cars respond to Uber-like requests for taxi rides. They tend to be cheaper than ordinary taxis and ordinary Ubers, because there is no driver to be compensated. Examples of robotaxi services in the United States are Waymo, Cruise and Zoox. Waymo began as the Google Self Driving Car Project. It now has operational robotaxi services in Phoenix, San Francisco and Los Angeles [16]. 

Photo credit: [P2]

In China, robotaxi service providers include Apollo Go, Didi, Huawei and WeRide. Apollo Go [17] operates a fleet of 500 vehicles in Wuhan, with fares markedly lower than those for a taxi driven by a human.

There are many examples on YouTube of successful robotaxi journeys.  These videos demonstrate the capabilities of autonomous cars to operate safely in the overwhelming majority of circumstances: to avoid collisions with other cars, cyclists and pedestrians, to turn and change lanes safely, to obey traffic signals, to negotiate one-way streets, and so on. 

Robotaxi services are essentially an application of artificial intelligence (AI), which plays a fundamental role in shaping the autonomous decision-making capabilities of vehicles, and in facilitating vehicles to adapt, learn, and improve performance over time [18].

Robotaxi services have the potential to be combined with dynamic carpooling, and could provide a pathway to reduce private car ownership. 

THE CANBERRA AUTONOMOUS CAR SIMULATION

The transport strategy of the Australian Capital Territory [19] refers to autonomous vehicles and their “potential to significantly change how we use vehicles … when used as shared vehicles on a subscription service they will create new travel opportunities for Canberrans and support car free lifestyles”. However, there has been no official modelling and costing of how such services might operate in Canberra. 

In 2014 a friend and colleague [20] developed a simulation [21] to model the use of robotaxi fleets to meet the transport needs of Canberrans. He updated the simulation in 2022. Some key features of his model were:

  • it was based on using a currently available EV (Tesla Model 3 with Full Self Drive) as the fleet vehicle
  • it assumed 34,000 robotaxis and 1300 chargers spread over 9 locations
  • it assumed a workforce of 2785 staff to support charging, cleaning, maintenance and administration
  • it made detailed, documented assumptions about costs, travel patterns, and electricity consumption. 

Some findings were:

  • apart from exceptions for tradies and specialist vehicles, the 34,000 robotaxis would be sufficient to provide the 1.1 million journeys required on a week day in Canberra – about 33 daily journeys per car
  • using an unsubsidised fare of $4.84 for a 10km peak-hour trip (and $3.08 off-peak) the fleet business would make a surplus of $125M per year
  • costs would be lower than that of privately owned cars and lower than the real costs of fixed-route public transport
  • with this size fleet, 95% of journeys can begin within 1 minute of being requested.

Comments

These fleets (and they may be a single fleet) could be government owned, or could be privately owned.  Either way, they can be regarded as “public transport”, even though they are not travelling on fixed routes.

Public transport in Canberra is heavily subsidised. If the robotaxi model were implemented and embraced by the community, there would be a very significant reduction in the cost to Canberra ratepayers. In my opinion, though, even if robotaxi services were implemented, fixed-route public transport, whether by light rail or electric buses, should still play an important role. In particular, it could support efficient express travel between Canberra’s town centres: Gungahlin, Belconnen, Civic, Woden, Tuggeranong and the future Molonglo town centre. 

An important part of the model relates to the average occupancy of the cars during peak hour journeys. The model uses an average occupancy of 2.4 passengers for fleet cars arriving at key commute destinations at peak times. A key issue is how to ensure that this peak hour occupancy actually happens. Even with a significant price differential to encourage peak hour car sharing, will commuter behaviour embrace “dynamic car-pooling” for these journeys to the extent necessary to achieve these average occupancies?  

In practice, if this model were implemented (perhaps during the next decade) it would be introduced incrementally, beginning with a small fleet to test user acceptance. There would be a complex transport ecosystem comprising active travel, privately owned cars, fixed-route public transport, traditional taxis and Ubers, and the robotaxi fleet. Some people, perhaps the majority, will continue to own a car, while others might relinquish an owned car and use a combination of subscription cars, active travel and public transport. Some might use a traditional taxi or Uber service for night journeys, and robotaxis at other times.

Suppose that public acceptance allowed the robotaxi fleet to grow to 10,000 autonomous cars. In this eventuality, something like 25% of Canberra journeys would be routinely made by robotaxi. If 25% of Canberrans felt sufficiently confident to relinquish their owned cars, the total number of cars in Canberra would fall from the current level of 340,000 to around 265,000 (including the robotaxi fleet).

OBSTACLES TO THE AUTONOMOUS CAR TRANSITION

Will autonomous cars become an established part of transport services during the next (say) 15 years? This is uncertain, in my view.  As we have noted, there are the beginnings of such services in selected cities in the US and China. And there are rapid advances in AI which are driving the development of autonomy. But there are also three types of obstacles (technical, legal and psychological) to these developments.

Technical obstacles

Despite the many examples of successful robotaxi journeys, safety and reliability errors do still occur. For example, in early 2024 Waymo issued a recall for its self-driving car software after two of its vehicles hit the same truck minutes apart [22]. A robotaxi operated by Cruise (a service owned by General Motors) severely injured a pedestrian, leading to the suspension of its permit in California, and both Cruise and Waymo were criticised by police and fire departments for inhibiting rescue operations [23].

Several commentators have drawn attention to the risk posed by hacking and unauthorized access to robotaxi fleets. If autonomous vehicles were to populate our roads at scale in the future, any malicious hacking could have catastrophic consequences.   

Clearly, as with air crash investigations, every robotaxi accident will be analysed and measures will be implemented aimed at preventing or greatly reducing the occurrence of future accidents with the same cause.  

Legal obstacles

In April 2024 Australia’s National Transport Commission released a consultation paper on Automated vehicle safety reforms [24]. A key issue raised by the paper is that, if a vehicle is controlled by an automated driving system, a human user will not be able to ensure that the vehicle is driven safely. Who will be blamed for an accident? The paper suggests that this responsibility should lie with a corporation (such as the company that operates a robotaxi service, or the company that licenses the driving system software) with “the right skills and capabilities to take responsibility for the safety of the automated driving system for its on-road life”.

Psychological obstacles

There are psychological obstacles to autonomy and also to car sharing. They include:

  • people feel less safe in an autonomous car, regardless of whether published safety data shows that travel with human drivers can, in aggregate, be less safe
  • media reaction to an autonomous car accident will be of a different scale to a similar accident where a human driver is involved. 
  • many people are likely to be averse to sharing a car with strangers, especially at night or in less populated places (in the robotaxi model, such people would have the right to be the sole occupant of a car, but they would pay a premium for this)
  • many Australians love their cars, want to own them, and store some of their possessions in them. Car ownership enhances their feeling of independence, even if an Uber or a robotaxi could respond to a booking within a few minutes.

WHAT SHOULD GOVERNMENTS DO?

Based on the above reflections, I have some suggestions for government actions to make transport more sustainable in urban areas. They are based on the assumption that governments will actually wish to encourage a transition to more sustainable transport, with greater use of active travel and public transport, and with a reduction in car ownership.

An example of a road duplication in Canberra where the funds could have been better spent on active travel infrastructure. Photo credit: [P3]

The suggestions are framed with Canberra in mind (so, in this case, they are recommendations to the ACT Government) – but many of them may also apply to other urban communities. They are:

  • in order to facilitate greater use of active travel, build and maintain properly connected, well maintained and safe cycling and pedestrian path networks, along with appropriate calming of suburban streets
  • in order to facilitate greater use of fixed-route public transport and to reduce transport emissions, electrify these fixed-route services and build more (partially electrified) park-and-rides
  • introduce more “on demand” public transport services (without fixed routes) for off-peak travel in the suburbs (these could be supported by smaller vehicles rather than large buses)
  • to fund the above improvements, stop duplicating urban roads except where essential for safety reasons
  • encourage employers to commend dynamic carpooling services such as Uber Commute to their employees
  • conduct local trials of autonomous vehicles on urban roads, leading to an initial pilot of a small robotaxi service, to gauge public acceptance
  • publish data on accident rates for autonomous vehicles in multiple jurisdictions compared with overall road accident rates, and encourage public discussion about autonomous vehicle safety.

SUMMING UP

Brent Toderian urged us to re-imagine our cities and suburbs as “compactable, walkable and complete communities” where most journeys are made by active travel or public transport, instead of communities where cars are regarded as the only viable way to get around. I have tried to imagine this transition for my city (Canberra) which is far from being a “compact city”, and for similar urban areas in Australia.

In such cities travel by car is far more convenient for several journey types, such as doing a weekly grocery shop for a family, or transporting children to a variety of events.

The choking of our cities with cars is a particular issue for peak hour commuting. Accordingly, it becomes critical that we maximise the opportunities to car-share for commute journeys that (for whatever reasons) cannot be made conveniently by fixed-route public transport.

Dynamic car-pooling services, supported by mobile applications and online platforms, would seem to offer the most promising way forward to realise car sharing on commute journeys. Governments should strive to find mechanisms to encourage the use of dynamic carpooling, and should do this with an eye to developments in autonomous vehicle technology and robotaxi fleets, which may become a reality in the 2030s. 

The development of robotaxis tends to generate strong reactions, mostly hostile but some supportive. It would be valuable if robotaxis were the subject of a calm and considered debate. On the one hand, robotaxis cannot and should not be introduced in the absence of community acceptance, and the obstacles to that acceptance (technical, legal and psychological) must be acknowledged. On the other hand, robotaxis could play an important role in facilitating dynamic carpooling which would see less cars undertaking commute journeys.  They could also encourage some car owners to see them as a highly affordable alternative to car ownership.

 

FOOTNOTES
 
[1] The author is a member of the ACT Branch of AEVA. At the time of writing he was also the National Secretary, but this article reflects the personal views of the author. It is not intended to represent AEVA policies.

[2] Brent Toderian. Better transportation in a 5 crisis world. https://www.youtube.com/watch?v=DtZYknxAp4Q

[3] A slogan on a T-shirt worn at the 2023 AEVA Conference.

[4] “Sustainable transport” is used here to mean transport that minimises negative impacts on the environment and on human health and safety, while also being affordable and convenient (so that it is actually used).

[5] This statement formed part of CANA’s submission to the Federal Government’s Transport and Infrastructure Net Zero Roadmap in July 2024.

[6] Bureau of Infrastructure and Transport Research Economics (BITRE). 

https://www.bitre.gov.au/publications/2023/australian-infrastructure-and-transport-statistics-yearbook-2023/road-related-revenue-expenditure

[7] Jane Jacobs. The death and life of great American cities.

https://trove.nla.gov.au/work/6761071

[8] Angie Schmitt. StreetsBlog USA, 10 March 2016.

https://usa.streetsblog.org/2016/03/10/its-true-the-typical-car-is-parked-95-percent-of-the-time

[9] Benjamin Kaufman. The Conversation, 17 May 2023.

https://theconversation.com/smart-moves-how-auckland-can-get-more-for-its-money-from-on-demand-public-transport-205389

AT Local is an on-demand, ride share service which uses a mixture of small vans and cars, including EVs.

[10] As an example of grocery delivery developments, see: Rachel Williamson. The Driven, 12 September 2024.

https://thedriven.io/2024/09/12/huge-ev-charging-hub-to-service-electric-grocery-deliveries-in-australias-biggest-city/

[11] This Wikipedia article reports that GoGet had 65,000 members in 2015, but does not cite more recent data:

https://en.wikipedia.org/wiki/GoGet

[12] Paul Smith. Australian Financial Review, 21 August 2024.

https://www.afr.com/technology/uber-scraps-105m-aussie-car-sharing-division-20240821-p5k47d

[13] Carpooling (Is it dead?)

https://www.reddit.com/r/Frugal/comments/v6vkml/carpooling_is_it_dead/

[14] Fulli, 25 September 2024.

https://www.fulli.com/en/news/carpooling-dynamic-economic-and-ecological

[15] Uber Commute.

https://www.uber.com/en-AU/blog/uber-commute/

[16] Waymo: where you can go.

https://waymo.com/whereyoucango/

[17] Kathleen Magramo. CNN Business, 22 July 2024.

https://edition.cnn.com/2024/07/18/cars/china-baidu-apollo-go-robotaxi-anxiety-intl-hnk/index.html

[18] Divya Garikapati and Sneha Sudhir Shetiya. Autonomous Vehicles: evolution of Artificial Intelligence and the current industry landscape. Big Data Cogn. Comput. 2024, 8(4), 42.

https://www.mdpi.com/2504-2289/8/4/42

[19] ACT Transport Strategy 2020, page 19.

https://www.transport.act.gov.au/__data/assets/pdf_file/0016/1613302/200601-ACT-Transport-Strategy_web.pdf

[20] Kent Fitch, who lives in Canberra. I had worked with Kent during 2006-2011 on the development of the National Library of Australia’s Trove service.

[21] Canberra autonomous car simulation.

https://canberraautonomouscars.info/

[22] Peter Valdes-Dapena. CNN business, 14 February 2024.

https://edition.cnn.com/2024/02/14/business/waymo-recalls-software-after-two-self-driving-cars-hit-the-same-truck/index.html

[23] Dara Kerr. NPR business, 30 December 2023.

https://www.npr.org/2023/12/30/1222083720/driverless-cars-gm-cruise-waymo-san-francisco-accidents

[24] National Transport Commission, 23 April 2024.

https://www.infrastructure.gov.au/have-your-say/automated-vehicle-safety-reforms

PHOTO CREDITS AND LICENCES

[P1] Photo by lorenz.markus97. https://creativecommons.org/licenses/by/2.0/

[P2] Photo by Bill Abbott. https://creativecommons.org/licenses/by-sa/2.0/

[P3] Photo by the author.