Articles about technology and the future of transportation rarely used to get far without mentioning jet-packs: a staple of science fiction from the 1920s onwards, the jet pack became a reality in the 1960s in the shape of devices such as the Bell Rocket Belt. But despite many similar efforts, the skies over our cities remain stubbornly free of jet-pack-toting commuters.
For a novel form of transport to make a material difference to our lives, several key requirements must be satisfied. Obviously the new technology must work safely, and operate within an appropriate regulatory framework. But public acceptance and solid business models are also vital if a new idea is to move from R&D lab to testbed to early adoption, and eventually into mainstream usage.
There’s inevitably a lot of hype surrounding the future of transportation, but also plenty of substance, with big investments being made both by disruptive tech companies and by incumbent industry players. Can technology help to get us and our goods around quicker, in greater safety, and with less damage to the planet?
Connected & Autonomous Vehicles (CAVs)
Driverless cars, or Connected and Autonomous Vehicles (CAVs) are getting the lion’s share of attention, but the wider implications of CAVs and other novel forms of transport are also firmly on the agenda — including smarter, greener cities and more efficient distribution of freight and consumer deliveries.
Most of the technologies listed here are in the early stages of the progression towards mainstream adoption, according to Gartner, with only five out of 30 making it beyond the Trough of Disillusionment.
No surprise, then, that there’s a lot of activity in the CAV market. In a report published last October The Brookings Institution collated reports of “investments and transactions attributable to autonomous vehicles or core technologies” between August 2014 and June 2017, and found over 160 separate deals amounting to some $80 billion. These covered auto electronics, microchips, rideshare apps, AI/deep learning, digital mapping, non-AI software, physical systems and sensors. The authors concluded that “investment in 2018 should be substantially more than the $80 billion disclosed from 2014 to 2017, and continue upward for some period of time as the race to deploy self-driving moves on.”
At the same time, public perception of autonomous vehicle safety seems to be heading in a positive direction. In a survey last year, Gartner found that while 55 percent of respondents (from the US and Germany) would not consider travelling in a fully autonomous car, 71 percent would ride in a partially autonomous vehicle.
These findings are echoed by the Deloitte 2018 Global Automotive Consumer Study, which found that the percentage of respondents considering fully self-driving vehicles unsafe ranged from 57 percent (in Japan) to 22 percent (in Mexico). In the previous year’s survey the figures were much higher, ranging between 81 percent (S Korea) and 54 percent (Brazil):
Still, as Deloitte notes, there’s a way to go when it comes to the perception of fully autonomous vehicles, with “almost half of consumers in most markets doubting the safety of this technology.”
Levels 3 and above are considered to be ‘automated driving systems’. In a Level 3 vehicle, the system handles steering, acceleration and deceleration, and monitors the driving environment, with human intervention available on request. A fully automated Level 5 vehicle does not require a steering wheel, pedals or any other controls — humans, if present, are simply passengers.
Many trials of autonomous vehicles are underway around the world, with the highest concentration in California, which not coincidentally provides the best statistics on traffic accidents involving them. Since 2014, California’s Department of Motor Vehicles (DMV) has logged 54 autonomous vehicle accident reports (as of 18 January 2018).
Reading through these reports, we judged that only four (7.4%) could be ‘blamed’ on the AV — and every one of those was under manual control at some point during the incident. Almost all were minor, low-speed accidents with no injuries, and the majority (56%) involved the AV being rear-ended by a vehicle driven by an inattentive human.
The California DMV is currently in the process of amending its regulations to allow the testing of fully autonomous vehicles without drivers (i.e. Level 5 vehicles).
Update (1 May 2018)
At around 10pm on 18 March a pedestrian in Tempe, Arizona was hit by an Uber Volvo XC90driving in autonomous mode (with a safety driver on-board) and later died of her injuries. Despite the array of on-board sensors, Uber’s vehicle failed to detect the bicycle-pushing pedestrian crossing the road, and video footage showed that the safety driver did not appear to be paying due attention to road conditions. Investigations are still ongoing, and in the meantime testing of Uber’s autonomous vehicles in Arizona has been suspended.
On 23 March the driver of a Tesla Model X was killed in Silicon Valley when, in Autopilot driver-assistance mode, the car collided with a concrete lane divider at high speed. Again, investigations are not yet complete, although a 30 March statement from Tesla had this to say: “The driver had received several visual and one audible hands-on warning earlier in the drive and the driver’s hands were not detected on the wheel for six seconds prior to the collision. The driver had about five seconds and 150 meters of unobstructed view of the concrete divider with the crushed crash attenuator, but the vehicle logs show that no action was taken.” Here’s a take on what may have happened.
Clearly, the interplay between autonomous systems and humans will be an issue for some time to come.
The key to improving the performance and safety of autonomous vehicles lies in the maturing of the underlying vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communications and (real-time) data processing systems. Over time, the components involved — including LiDAR systems, for example — will become cheaper, helping to remove another barrier to adoption. They will also become more power-frugal, making them more suitable for deployment in cleaner electric vehicles.
Autonomous vehicles are likely to prove much safer than conventional ones, and public confidence in them will surely increase. But even in the best-case scenario for CAVs, there will inevitably be a time lag as conventional vehicles are gradually phased out: the average age of light vehicles in the US in 2014 was 11.4 years, for example. The transition period, which could last well over a decade, will see roads carrying a mixture of autonomous and human-driven vehicles, and is likely to prove challenging for drivers, passengers, regulators and enforcement agencies.
And of course, once CAVs become the norm they will still have to deal with pedestrians and other unpredictable non-vehicular elements in the environment. There may be fewer auto accidents, but insurance companies and lawyers will doubtless still find a way to make money.
Most people are familiar with drones from footage in the media, and low-end devices are now affordable enough for enthusiasts to get directly involved. Meanwhile, on the commercial side, logistics companies like Amazon (Prime Air), DHL and UPS are investigating the use of drones for parcel distribution — particularly ‘last mile’ deliveries in rural areas where conventional vans and trucks can struggle. Google’s X ‘moonshot factory’ is also doing R&D on delivery drones under its Project Wing.
As far as public acceptance is concerned, the position on drone deliveries seems to be ‘interested but wary’. An online survey conducted by the US Postal Service in June 2016 found that while three-quarters of the 1,465 respondents expected drone delivery by 2021, less than half (44%) liked the idea. Drone malfunction was the main concern (46%) with theft (16%) and intentional misuse (14%) much less serious worries. Speedy delivery was the main reason for interest in the technology, with emergency delivery also highly ranked. In the UK, the IMRG Consumer Home Delivery Review 2016 found that only a quarter of its 1,280 survey respondents (25.6%) would be prepared to have parcels delivered by drone — up slightly from the previous year’s survey (23.8%).