Autonomous vehicles and the changes ahead
Social and economic impacts
Widespread adoption of highly autonomous vehicles would transform the global marketplace. Millions of new jobs and immense economic opportunities would emerge. But people and companies relying on transportation industries would lose out.
For example, automated trucking could unlock exciting safety and economic benefits for some. It would reduce the impact of driver fatigue in crashes. With improved logistics consumers would get their stuff faster. But for this change to occur – millions of human truck drivers would lose their jobs. Massive economic flow-on effects would impact businesses supporting truck drivers. Motels, rest stops and local mechanics could all suffer too.
This is an ongoing concern across all industries facing the transition to increased automation. Policy researchers are calling for governments to understand and plan for negative impacts. A short-term fix might involve funding education and retraining programs for displaced workers. An example long-term solution could be the implementation of a universal basic income.
There are also thoughts that the institutional knowledge of drivers and logistics professional will add a lot of value to aiding automated freight and logistics vehicles operate safely and efficiently.
The change that is coming the way of this transport sector is something iMOVE is investigating in its Workforce implications of transport digitalisation and automation project.
Vehicle ownership and personal identity
Automation enables a range of transport options that may not have been possible in the past. It will also likely mean a change in the way society thinks about vehicle ownership in the future.
Individual vehicles owners may become far less common. Vehicle fleets may end up being owned and operated by private or public entities. This will likely meet resistance – due to the strong sense of identity and enjoyment some people associate with car ownership.
Vehicle ownership / use and traffic congestion
Vehicle ownership, and vehicle use, also has a strong bearing on hopes held that automated vehicles will help reduce traffic congestion.
On the positive side of looking at the introduction of autonomous vehicles on public roads, they may reduce traffic congestion in two key ways:
- They will ideally increase road safety and reduce crashes that result in significant traffic delays.
- They could reduce start-stop traffic created by human drivers – allowing for more efficiency even in congested conditions.
However, if ownership of autonomous vehicles replicates the current level of private vehicle ownership, there will be minimal change to our current congestion woes.
Worse still is the issue of what an autonomous vehicle does while it is not being used for trips with passengers. Does it park? Or does it drive around in circles awaiting the task of picking up its passengers? Clearly this behaviour would ramp up traffic congestion.
And there is some evidence on the negative influence of driverless vehicles on traffic congestion. In a Massachusetts Institute of Technology study, Impacts of transportation network companies on urban mobility, on the affect of Uber and Lyft vehicles in New York City. The result since the introduction of these services saw road congestion in terms of intensity increase by 0.9%, by duration a 4.5% increase, accompanied by an 8.9% decline in public transport use and an insignificant change in vehicle ownership.
Admittedly these figures stem from the addition of shared vehicles rather than autonomous vehicles, but they go some way to illustrating the possible consequences of change.
“It is far from clear how much of the congestion challenge can be resolved through greater sharing of private cars (no matter whether they are autonomous or not), increasing occupancy, assuming a constant number of person trips. However, sharing of private cars could lead to increased trips overall through a higher number of trips per vehicle, and to greater congestion if the number of trips overall goes up.” Says Professor David Hensher in Smart shared mobility and potential implications for levels of congestion.
Operational design domains: The behaviour of autonomous vehicles
The Operational Design Domain (ODD) articulates the most favourable conditions for autonomous systems and features to operate safely. It guides how autonomous vehicles are designed, built, tested and regulated.
Features intended to support Level 1 or 2 automation can only operate under very specific conditions. Their ODD is limited to those conditions and operating outside of them would be unsafe. For example, intelligent parking systems can only take over if your vehicle is close to a suitable parking space.
Level 5 vehicles would be able to operate under all driving conditions. Their ODD would be infinitely more nuanced and complex. A vehicle operating at this level would be guided by an ODD that defines conditions for a multitude of features, systems and processes.
Some factors the ODD considers:
- Types of roads: for example, will features that support automation on highways or residential streets?
- Topography: for example, will the vehicle operate on hilly terrain, flat urban roads or desert environments?
- Speed constraints: low-speed autonomous shuttles typically operate at speeds below 15mph.
- Ecological circumstances: for example – is it safe to operate the vehicle in rain or snow? Also does a particular feature perform better in daylight or at night?
- Other specialised domain constraints: might include roadside infrastructure, traffic signs, laws and safety risks associated with a particular location or use case, and the co-existence of autonomous vehicles with legacy (non-autonomous) vehicles.
Ethics
Human drivers make life and death decisions every time they drive. Autonomous vehicles will need to have the same ability for critical decision-making.
Ethics advocates believe it’s important to build ethics into the technology from the ground up. Thought experiments can help us to understand the complicated nature of this issue:
Imagine you are driving your car on the highway. You see something large fall off a truck in the lane ahead of you. You must swerve to avoid it. To your left is a family car and to your right is motorcycle with a single rider. Do you swerve left or right, or do you impact with the object on the road?
We may be able to accept a human’s imperfect reactions and decisions when collisions occur. But autonomous vehicles will be held to a higher standard. What types of objects should the vehicles crash into or avoid? Who should be making these decisions?
Autonomous vehicles’ decision-making algorithms are programmed by people. Are the humans who write decision-making algorithms responsible for fatal crashes? And, how to we avoid hard coding inherent human biases into software when the stakes are this high?
Laws and regulation
Policymakers and governments are already working to understand and regulate automated vehicle safety. The work of updating local, state and national laws to support widespread adoption will take years. Currently, many countries allow on-road trials under controlled conditions, including Australia.
Regional differences in laws may complicate roll-out of autonomous vehicles for global companies. Local laws and vehicle standards will mean a ‘one-size-fits-all’ approach will not work for manufacturers. Roll-out may be further delayed in some countries if their laws vary significantly between local jurisdictions.
Adding to this challenge is the fact that technology continues to evolve. This means policymakers and governments are writing laws to regulate vehicles that don’t even exist yet.
