Vehicle automation is expected to have a huge impact on demand for energy in transport, but there is very little clarity as to what the net effect will be.
A review of literature examining the issue -- Autonomous Vehicles and Energy Impacts: A Scenario Analysis, published in Energy Procedia in December 2017 -- found that the various studies undertaken pointed to anything between a 64% fall in transport energy consumption by 2050, compared with 2017, to a 205% increase.
This wide spectrum of outcomes relates to total transport energy consumption, not to the source of that energy, which was outside the scope of the study.
As a result, to determine changes in future oil demand, analysts must assess the impact of competing fuels and energy sources – biofuels, electricity, hydrogen, natural gas/LNG – and then place that analysis within the context of the behavioural changes brought about by automation.
A 50% share of electric vehicles (EVs) in the passenger vehicle fleet by 2050, for example, looks very different in the context of a 64% fall in overall energy transport demand – implying a potentially radical drop in oil demand -- when compared with a 205% increase, in which the impact of passenger vehicle electrification may be entirely negated.
There is nothing intrinsic to automation that implies transport electrification, but there are clear synergies.
For example, while recharging time is considered a negative feature of EVs, compared with internal combustion engines (ICE), it would be a much safer means of reenergising a vehicle in the absence of a driver. In addition, vehicle sharing, a development expected to result in fewer vehicles travelling more miles, would favour the lower fuel and maintenance costs of EVs.
What is apparent is that the first generation of robo-taxis is set to be almost exclusively electric or plug-in hybrid.
Volvo intends to supply the 360c to Uber, Waymo is building its autonomous fleet with plug-in hybrid Chrysler Pacific Minivans and all-electric Jaguar SUVs, Navya’s Autonom CAB is electric, while Cruise Automation will use parent company General Motors Chevrolet Bolt. BMW is looking to its iNext EV and nuTonomy is moving from Renault Zoes and Mitsubishi i-MIEVs to the Peugeot 3008, which has both ICE and electric drive train options.
The worst possible combination of scenarios, from an oil producer perspective, would be a fall in overall energy demand with the development of automation and electrification proving highly symbiotic. High levels of automation could accelerate electrification. Alternatively, a high level of automation combined with drive train neutrality, particularly in the commercial vehicle sector, offers the highest oil demand outlook.
The different outcomes depend upon the weight given to the possible changes in driver behaviour and vehicle use.
Automation could increase average travel speeds as a result of better traffic conditions, resulting from fewer accidents and improved traffic flow. The number of drivers could increase as both youth and old age are reduced as barriers to driving, while people prevented from driving as a result of disability would also benefit.
A further possibility is that an improvement in ‘in-car value time’, where the driver (now a passenger) can work via full interconnection to the internet, or indulge in some form of entertainment. In-car connectivity and automation could make point-to-point car travel preferable to mass transport even for longer journeys, challenging short haul flights and train journeys.
As a result, automation could both increase demand for travel and cause a switch in travel modes from mass to individual transport. This latter development could also mean a switch from oil product use (jet fuel and diesel) to electricity.
On the other side of the coin, automation may promote eco-driving, with the car’s computer prioritising fuel efficiency above all else.
This possibility is extended with platooning, whereby fully-automated trucks drive in convoy from major distribution hub to major distribution hub, cutting down on wind resistance and driving at optimal speeds for saving fuel. Right-sizing and car-sharing are other means by which vehicle miles travelled (VMT) could be cut.
The study concludes that full automation will “induce travel demand and attract new user groups, which will generate more trips and VMT that results in more energy consumption.” However, dynamic ridesharing and shared autonomous vehicles are seen as means of reducing the impact on transport energy consumption, particularly in urban environments.
While the increase in driver pool, improved safety (eventually) and overall increase in attractiveness of road transport provided by full automation look relatively straight forward, the idea of shared vehicles implies a more radical and uncertain change in consumer behaviour.
Private car ownership has never been a commercial proposition but one of convenience. Privately-owned vehicles are heavily under-utilised capital investments that spend most of their time doing nothing.
The shared vehicle concept assumes that robo-taxi costs fall so low that travellers give up the convenience of owning their own car. Unlike its private counterpart, the shared car is highly utilised – it will need to be to keep costs competitive with both public transport (where it can offer better point-to-point journeys) and private ownership.
However, there are questions over whether robo-taxis could meet both the need for high utilisation and the same level of convenience offered by private ownership.
Full automation also increases the convenience of private ownership. Moreover, transport demand is not constant, but has peaks and troughs largely centred around commuting times. To deliver convenience, sufficient robo taxis would be needed to meet peak demand, but that implies less utilisation during other periods.
This assumes that robo taxis are organised as commercially-owned fleets, which appears to be the direction of travel backed by joint ventures combining technology firms, ride hailing platforms and auto manufacturers. This means a major change in business model and cost structure for companies like Uber – capital investment in fleet ownership and maintenance, as oppose to sourcing work for privately-owned, driven and maintained assets.
While robo taxis will make taxi drivers redundant, it is not clear that they will make taxi travel sufficiently cheap and convenient to displace privately-owned car ownership or compete with public transport to the extent expected.
The wide range of possible outcomes will also be affected by externalities. An increase in VMT could work against improvements in traffic flow and potentially negate faster average speeds, if not accompanied by increased investment in roads.
Given the uncertainties, it is difficult to come to hard and fast conclusions at this stage, but the balance of probabilities suggests that, first, given the uncertain impacts of vehicle sharing, automation is more likely than not to increase overall demand for transport, most likely on a post-2030 timescale. Second, automation and electrification might be complimentary, but they are not co-dependent. Automation could therefore advance in the heavier commercial vehicle sector based on ICE technology.