How Can We Effectively Deliver the Benefits of Automated Driving Technologies Over the Next Decade?
Has the autonomous revolution stalled?
In the eyes of the global media, the answer to this question is ‘yes’. Over the last 5 years, the enormous hype surrounding automated vehicles and the incredible benefits to be gained – safety, efficiency, equity – have slowly retracted to a state of realism.
Uber, one of the biggest AV players and beneficiaries, recently decided to terminate plans for AV taxis, selling off its autonomous division (ATG, or Advanced Technologies Group, pictured left) to Aurora in a deal worth about $4bn. This is a significant market signal from one of the industry pioneers who were investing close to $20mn per month for the last few years and acknowledges the huge challenges of achieving this ‘moonshot’ future.
But the technology quietly continues to progress
Around the noise of Uber’s announcements and beneath a veil of realism, the industry quietly grinds on.
In the US, Waymo continues to make progress and last month (with little fanfare) revealed it had passed a major milestone and opened their autonomous taxi service to the public in Chandler, Arizona. In recent months, Zoox and Cruise have both unveiled purpose-built vehicles for ride-hailing services without a steering wheel or pedals. May Mobility has also announced a partnership with Via that will enable their AV shuttle vehicles to be integrated with public transportation systems.
In the UK, Oxbotica recently raised $47bn to accelerate the development of ‘universal autonomy’ – the flexible software needed to run self-driving vehicles in multiple environments, whilst Aurrigo has been deploying autonomous vehicles at golf courses, airports, and universities.
This is a significant progression, that shows commercial applications of the technology are reaching a tipping point.
And, importantly, the regulatory environment is catching up
At the Federal level in the US, legislation has been slow to address the legal risks. The National Highway Traffic Safety Administration (NHTSA) has offered broad and voluntary guidance – not wishing to stifle innovation in this valuable industry. But in December 2020, the NHTSA began soliciting public comments to standardize safety protocols for driverless vehicles. This an important step towards enabling large-scale AV deployments over the next few years.
Irrespective of Federal guidance, individual states have been making major strides towards an autonomous future. By the end of 2020, 45 out of 50 states have enacted AV legislation. Once the overarching safety standards are in place, the US is set to accelerate the deployment of AVs.
In the UK, the Government has set its intention to “lead the way globally in embracing the safe development of driverless technology”. The Automated and Electric Vehicles Act 2018 (AEV Act) received Royal Assent in July 2018 and more than 90 testing and research projects have now been funded to better understand the opportunities and challenges which exist.
With testing projects ongoing in cities including London, Coventry and Oxford (including our own Project Endeavour, pictured left), the updated Code of Practice for Automated Vehicle Trialling provides further clarity as to what is expected of organisations wishing to trial autonomous vehicles on the roads.
Further reform to existing legislation and the Highway Code is under consideration to allow the use of ‘Advanced Lane Keeping Systems’ (or ALKS) that can take control of vehicles and can keep them in the lane on motorways. This significant regulatory advancement can accelerate the rollout of semi-autonomous technologies.
So we must focus on enabling the high-value, low-barrier use cases
The Society of Automotive Engineers’ Level 1-5 taxonomy (pictured right) is valuable to define the technological advancement of automated vehicles but to understand the early commercial applications, a use-case-based approach is more helpful. This involves the specific understanding of what is being transported, how this is being transported (i.e level of autonomy), where are they operating, who is operating the vehicles and when is this likely to be viable – and the associated risk of deployment.
What is being transported?
The first element of the use-case-based approach is what is being transported? The level of risk and opportunity for early deployment is different depending on whether we’re transporting people or different types of goods.
How is it being transported?
Or under what level of autonomy. Beyond level 2, the vehicle becomes ‘self-driving’ when automated features are engaged. Level 3 enables traffic jam support with the potential to hand back control, whilst level 4 enables full self-driving capabilities under limited conditions and level 5 can drive anywhere in all conditions.
Where are they being transported?
The operational environment is a huge factor in the development timeline of a self-driving car – dictating every key technical challenge. Requirements for hardware and software are different for driving in urban environments versus highway-type environments versus controlled campus environments.
Who is transporting it?
Is it going to be individuals, or are they going to be owned by private fleets? This is important, since the operating models and the economics are going to be different, but crucial to ensuring viability of a new service.
When is it likely to be available?
Timing is everything for this industry, amid a barrage of broken promises throughout the 2010s from the automotive and technology industries. The applications with the least technological, political, legislative, social, and commercial barriers are likely to be rolled out earlier, whilst some of the ‘moonshot’ opportunities for automated vehicles may be pushed back another 10-20 years.
And early commercial deployments are accelerating
In crisis times such as a pandemic, the value of safely moving critical goods significantly increases, so we are seeing accelerated AV deployments.
In the trucking sector, TuSimple is starting to sell their technology to logistics operators, enabling fully autonomous cargo movements from hub to hub over US highways, promising real savings in cost and time.
At the local scale, Nuro is set to be California’s first driverless delivery service and will be carrying groceries and medicine to local communities.
If we are moving people, the environment becomes more important. Controlled, campus environments carry a lower risk than uncontrolled, mixed use environments.
Back in 2018, we saw US-based AV start-up, Voyage, secure an exclusive, multi-year license to deploy an autonomous ride-sharing service at the largest retirement community in the world in Florida. This is a simpler and slower environment for AVs that can still benefit thousands of people.
These are both examples of lower risk, high-value deployments of the technology that are facilitated by a simpler operating domain.
But we must utilise modelling and simulation
This rapidly emerging market requires a sophisticated approach to decision making, especially if the potential benefits are to be fully realised.
But how do you make such critical business decisions when you’re innovating into the unknown with different environments? Black box AI-based optimisation tools don’t fully explore and uncover the different trade-offs you’re making, whilst data analytics look back at historical trends that aren’t useful for assessing disruptive futures.
The answer lies in simulation-based scenario analysis.
When you combine ‘what-if?’ style questioning with city-scale simulation, you can test possible deployment scenarios in many different environments and explore trade-offs in order to manage risks, respond to opportunities and ultimately find the best solution for cities, citizens and your business.
Delivering the Benefits of Automated Driving Technologies
Beyond the negative media headlines across the autonomous vehicle industry, there remain near-term opportunities to use these technological advancements to benefit society – which does not necessarily need to involve robotaxis.
Increased regulation, vehicle testing, and pilot programmes are slowly building trust in these systems, and by focusing on the lower risk, high value use cases based on the understanding of what, how, where, who and when required, we can begin to socialise these vehicles in society.
Before deploying vehicles in commercial applications, we must use modelling and simulation tools to test the business model strategy and configuration of vehicle service – in order to deploy the most effective services for our communities.
To learn more about applying scenario-based simulation in your autonomous fleet strategy development, book in for a free pre-consultation here.