Autonomous Vehicles: Making the improbable, possible.
In 2004, two of the smartest guys around, Professor Frank Levy from MIT and Professor Richard Murnane from Harvard, published a book called The New Division of Labor: How Computers Are Creating the Next Job Market.[i] They asked the question “What kind of tasks do computers perform better than humans?” Their conclusion was that it continues to excel at rule-based logic, but struggles with tasks that involve expert thinking and complex communication.
They suggested that computers, while disruptive and impressive, would not be able to replace humans in jobs that required a variety of observational and decision-making skills. The example they gave was executing a left-hand turn in oncoming traffic. “As the driver makes his left turn against traffic, he confronts a wall of images and sounds generated by oncoming cars, traffic lights, storefronts, billboards, trees, and a traffic policeman. Using his knowledge, he must estimate the size and position of each of these objects and the likelihood that they pose a hazard.”[ii]
They therefore concluded it was extremely unlikely that we would see autonomous vehicles on the road anytime soon. But just six years later, Google’s first-generation autonomous cars were doing just that. Then, two years later, on October 23, 2012, California Governor Jerry Brown sat alongside Sergey Brin from Google and signed a new bill that allowed autonomous cars on the state’s roads.[iii] “Today we’re looking at science-fiction becoming tomorrow’s reality—the self-driving car,” Brown stated.
Since 2012, the buzz around autonomous vehicles has reached fever pitch. Initially pioneered by Google who modified a fleet of Toyota Priuses, now nearly all the major car companies have advanced self-driving car projects in the works, and one company in particular – Tesla – already provides semi-autonomous capability via its Autopilot hardware that was installed in every car coming off the production line since October 2014.
How do they work?
The US Department of Transportation defines full self-driving autonomy as the vehicle being designed to perform all safety-critical driving functions and monitor roadway conditions for an entire trip. The vehicle is equipped with sophisticated sensors that include radar, LIDAR, sonar and multiple, high-resolution cameras to provide object detection, and actuators that control the electric steering, braking and throttle control to provide collision avoidance if obstacles are detected. All controlled by complex and sophisticated software.
Elmar Frickenstein, Senior Vice President of BMW, explains the level of end-to-end architecture required to develop a truly autonomous car. “Inside the car, we have to have a central computing platform — a supercomputer provided by [Intel]. In addition we need motion control for highly dynamic and elegant driving, a sensor cluster and a safe, secure and private environment. Outside the car, we need a high-performance cloud solution to help us with things like swarm intelligence for new driving strategies, planning, data analytics and machine learning to create new functions.” GPS will provide route and traffic information to its destination so the car knows where it’s going and how it’s getting there.
While autonomous cars are getting most of the media attention, the same technological advances are being used to make a whole variety of vehicles capable of operating without a human driver. These range from the very large – such as ocean going sea vessels being developed by Rolls Royce, autonomous cargo planes being developed by Airbus and autonomous trains being developed for Rio Tinto, to the very small, such as pavement delivery robots such as those being developed by Starship and autonomous drones, first demonstrated by Jeff Bezos of Amazon.
Autonomous cars are being developed by nearly all the major manufacturers, specifically Ford, Daimler, BMW, VW Group and Toyota, plus new entrants such as Tesla. However, technology companies such as Uber, Lyft, Waymo, Apple and Google and Amazon are all working on autonomous vehicles.
The same technology can of course be used in larger vehicles such as trucks. Existing manufacturers such as Volvo and Daimler, logistics companies such as FedEx, UPS, US Postal Service and technology companies such as Otto and Tesla, all intend to have their versions of the driverless truck on the road in the next two or three years.
There are variations of autonomous capabilities from fully autonomous vehicles to platooning, where a lead vehicle with a human driver has a convoy of other vehicles behind it, all connected and autonomous copying and following the actions and speed of the lead vehicle.
There have even been examples where different autonomous capabilities are combined – for example Daimler is partnering with Starship robots on the development of a mothership van concept. An autonomous van loaded with Starship delivery robots travels to an optimal location, the delivery robots scatter from the van and travel to each house, deliver the goods and then return to the van to be reloaded with the next batch of orders.
When joined together, autonomous vehicles are capable of disrupting almost every stage of the end-to-end supply chain. For example:
- autonomous mining vehicles will dig material out of the ground and…
- move it to an autonomous train that drives the goods to…
- a fully autonomous port, such as the one in Rotterdam, loads goods onto…
- an autonomous ship, which travels to another autonomous port,
- unloads the goods onto an autonomous truck, which…
- drives to a factory, where the goods are used in the production of finished products, that are…
- loaded onto another autonomous truck and driven to a fulfilment centre where…
- they are picked and packed by robots before ‘the last mile’ is completed by…
- autonomous drones / robot or delivery vehicle, delivering the goods to the end consumer.
Autonomous vehicles: a compelling business case
Firstly, they stand to reduce the number one cause of vehicle accidents – humans. Human error is responsible for 70-80% of all maritime accidents[iv], 90% of all haulage accidents and 95% of all car accidents (car accidents alone were responsible for 1.25 million deaths in 2013).[v] Widespread adoption of self-driving vehicles could eliminate around 90 percent of all auto accidents in the US, preventing up to $190 billion in damages and health costs annually. This will cause a productivity boost due to saving time normally lost through recovering from accidents.
Secondly, the biggest cost in nearly every vehicle-related activity is usually the human at the front. Take the driver out and the costs reduce significantly. As Uber’s ex-CEO Travis Kalanick declared in 2014: “You’re not just paying for the car; you’re paying for the other dude in the car. When there’s no other dude in the car, the cost of taking an Uber becomes cheaper than owning a vehicle.”[vi]
Research by Ark Invest[vii] highlights that the average cost-per-mile of travelling in a US taxi is $3.86, whereas a San Francisco Uber is $2.25 per mile, and travelling in your own vehicle costs $0.76 per mile. However, the estimated cost of travelling in a shared autonomous vehicle in 2020 is predicted to be just $0.35 per mile. Morgan Stanley has estimated that autonomous technology could save the freight industry upwards of $168bn annually, $70bn of which would come from reducing staff, $35bn from fuel savings, $36bn from accident savings and $27bn from productivity gains.[viii]
Thirdly, commuting consumes an enormous amount of time that could be used more productively. Management consultants McKinsey estimate that worldwide time saved every day by driverless cars could total as much as one billion hours. Morgan Stanley has estimated that autonomous vehicles will generate $507 billion in productivity gains in the US.[ix]
Finally, human drivers are limited by their biological constraints, needing to rest, take breaks and sleep. Autonomous vehicles need none of this, meaning use rates will rise dramatically. When productivity gains are combined with other savings such as fuel efficiency and accident avoidance, Morgan Stanley estimates the total potential annual savings to the US economy will be $1.3 trillion per annum. Worldwide, that number could reach $5.6 trillion.[x]
Safer, cheaper, harder working – what’s not to love? Unless, of course, you are currently employed as a driver. According to the American Trucker Association, there are 3.5 million professional truck drivers in the US and an additional 5.2 million people in non-driving roles in the industry. That’s 8.7 million trucking-related US jobs that could potentially be eliminated in the next five to ten years due to the impact of autonomous trucking alone.
Where next: future developments
While we are yet to see autonomous vehicles on the road in any great capacity, all this will change very soon. By 2020, autonomous taxis and trucks will be a feature in many countries such as the US and Japan, and by that time, autonomous road robots and drones should be a familiar sight.
Waymo has already announced that it will begin offering an autonomous taxi service in 2017, a Norwegian container ship, the Yara Birkeland, will be the world’s first electric, autonomous, zero-emissions ship when it launches in 2018, autonomous trucking has already started in the US, Starship road robots have been tested in London and Barcelona, and drones have already completed successful trials around the world. In many cases it is the legislation, not the technology, that is the barrier.
Most autonomous vehicles are also likely to be electric. Tesla CEO, Elon Musk, believes that half of all car production will be electric by 2027, and almost all will be autonomous.[xi] This will have a positive effect on the environment. The use of electric autonomous taxis alone could reduce greenhouse gas emissions by 87 to 94 percent per mile by the year 2030.[xii]
Autonomous vehicles are also creating significant opportunities for rethinking business models. Today the car industry is measured by the number of cars sold; tomorrow it will be measured by how many passenger miles it undertakes and the margin per mile. As cars become autonomous, the business model will move towards access, and away from ownership. The days of owning a rapidly depreciating car that is only utilised 5% of the time are coming to an end. A Jetsons future awaits.
[i] Frank Levy, Richard J. Murnane, ‘The New Division of Labor: How Computers are Creating the Next Job Market’, p28 2004, Princeton University Press.
[ii] Frank Levy, Richard J. Murnane, ‘The New Division of Labor: How Computers are Creating the Next Job Market’, p28 2004, Princeton University Press.
[iii] Bloomberg. Driverless Car Future Sends Google, U.S. to Figure Rules By Angela Greiling Keane – Oct 24, 2012
[ix] ‘Autonomous Cars: The Future is Now’; Morgan Stanley Blue Papers, Jan 23, 2015 https://www.morganstanley.com/articles/autonomous-cars-the-future-is-now
[xii] Philip Oldfield; ‘Electric ‘robocabs’ would reduce US greenhouse emissions by 94% – study‘, The Guardian, 6th July 2015 https://www.theguardian.com/environment/2015/jul/06/electric-robocabs-reduce-us-greenhouse-emissions-94-percent-study