Transport tech: four amazing vehicles that could change the world
In 1908, automotive pioneer Henry Ford changed the world with the release of the Ford Model T. The new technology changed not just how we travel, but how we live and work too. A century on and it's starting to feel like we're approaching a similar moment again.
Several new technologies are on the cusp of breaking out from the test track and going mainstream. In the not too distant future, we expect that our cars will all be electric. And what about drones?
Will we soon no longer impatiently wait for a delivery van, but instead look to the heavens? Some of this might sound like science fiction right now, but the transport revolution could be closer than we think. So how close are we to the next Model T moment?
Read on to find out. Which? Tech Support - show your devices who's boss
Thanks to the same artificial intelligence (AI) and machine learning technology that has revolutionised everything from Google's search results to how banks spot card fraud, we really are closer than ever to stepping into a car and not paying attention while behind the wheel. The idea is simple: instead of trying to program into computers what traffic lights look like, or how to spot an oncoming lorry, it's much easier to teach them to spot patterns. If you show an AI system thousands of photos of, say, a stop sign, it will figure out the sign's unique qualities and will be able to spot them itself with an impressive degree of accuracy.
The experimental phase
Most of the major car manufacturers, as well as the big tech companies, now have teams dedicated to developing autonomous technology.
There are different experimental projects taking place all over the world. For example, Ford is testing autonomous cars on the streets of its native Detroit, and if you take a trip to Palo Alto, California, you can take a ride in a fully autonomous Waymo taxi that is being developed by Google. In Britain, construction has just begun on a 180-mile test road near Birmingham, which will enable carmakers to test their driverless vehicles in realistic British conditions.
Collecting the driving data
Perhaps the most important unresolved question is exactly how autonomous cars should collect data for AI to analyse.
Most manufacturers are using a technology called Lidar, which works similarly to radar, but instead of radio waves it bounces a laser off objects to build up a detailed 3D image of the environment. The problem is that Lidar is still very expensive specialist equipment, which will ultimately make driverless cars more expensive, too. By contrast, Tesla is betting something more affordable: ordinary optical cameras.
It believes that accurate navigation can be achieved simply by having its AI algorithms analyse standard flat camera images. And if it can make it work using these standardised parts, it could reduce the cost considerably.
A gradual arrival
In reality, autonomy will appear on different types of transport and on different types of road at different times. In terms of technology, many cars already on the road contain some of the basic building blocks of autonomy, such as on-board cameras.
Some even have limited autonomous capabilities, such as adaptive cruise control that watches the car in front and matches its speed, or steering controls that will track the road and turn the steering wheel without the need for a human driver to do so.
2. Driverless HGVs
So where will we first start to see driverless vehicles? Surprisingly, it might be in the form of heavy-goods vehicles on motorways.
At first glance, this might sound incredibly dangerous. But counterintuitively, motorways are perhaps the safest place to do it, as they have something that smaller roads lack: predictability. The range of possible movements, hazards and behaviours is significantly more limited on a motorway compared with other roads.
There's only a handful of fixed entry and exit points, all vehicles are heading in one direction and all must keep moving. These are perfect conditions for training an AI to spot patterns: instead of having to teach the AI to spot thousands of different hazards and behaviours, it only has to spot a relatively small number, which in turn makes the AI pattern recognition much more reliable.
But why HGVs?
Here there is a more compelling business and usage case. The lorries that spend their days rumbling up and down the motorway tend to travel between a limited number of places, such as between a port and a warehouse.
Ports and warehouses are often located for obvious reasons close to motorways. Tests using HGVs have already taken place around the world, and last year a group called Helm UK had been due to begin a limited trial of the technology on British roads transporting real parcels for courier firm DHL. The launch has been delayed by the pandemic, but the plans are exciting: the idea is that it would trial not just autonomy, but a concept called 'platooning' where a human would drive one lorry and then would be followed by two autonomous lorries directly behind, saving person-power and, thanks to the physics of slipstreaming, fuel too.
Electric cars today are roughly where smartphones were around 2007, just after the iPhone launch. Most customers still remained unconvinced, even if they could afford to pay the high prices the new technology demanded. The stats on electric cars reflect the same attitude now.
Of all new cars sold today in Britain, just 4.7% are pure electrics such as the BMW i3 or Renault Zoe. But while this is a small slice of the pie, the numbers are growing. As of July this year, there were 136,600 pure electric cars on UK roads, and roughly the same number again of hybrids, such as the Toyota Prius, which use both an electric motor and an old-fashioned petrol engine.
So what's the hold up?
One concern is obviously cost.
The most expensive part of an electric car is the battery. But here, the technology is rapidly improving. According to one report by the consultancy McKinsey, between 2010 and 2016, the cost of batteries fell by 80%, from around £1,000 per kilowatt hour to £227.
One oft-voiced concern about electric vehicles is an obvious one: range anxiety. Here, the technology has also progressed a lot in the past decade. In 2010, the Nissan Leaf, widely viewed as the first mainstream electric car, could only manage about 71 miles before needing to be recharged.
The 2019 model, by contrast, can travel over twice that distance, clocking in at 151 miles, according to America's regulator. Recharging, however, is still an issue. With a petrol car, if you run low on fuel you can pull into any one of 8,000 petrol stations scattered across the country and be driving away again in minutes.
But for electric vehicles, while there are now upwards of 33,000 charging points across the UK, the act of charging is nowhere near as fast.
Charging technology is slowly improving, but we're unlikely to reach petrol-pump style charging improvements any time soon. Instead, electric advocates prefer to make the point that electric vehicles are not just about a technological change, but a change in attitude too, that treats your car more like your smartphone. For most people on most days of the year, the argument goes, the batteries in your phone and your car will last you all day.
But what of occasional days when we want to take a longer trip? Just like we have to deal with the minor inconvenience of carrying a charger with us, with our cars we'll have to get used to the minor inconvenience of stopping for a little while to charge up mid-journey.
The best electric cars are every bit as good as their petrol or diesel rivals. To see which models have impressed our experts, head over to our guide on the best electric cars for 2021.
According to Transport for London, 13% of the capital's road traffic is light goods vehicles - 7,300 vans an hour during the morning peak period. So unsurprisingly, other delivery solutions are being sought. This problem is obviously of interest to Amazon - the origin of many of those deliveries in the first place.
That's why for several years the company has been working on the technology to turn drone delivery into a reality. Amazon's so-called 'Prime Air' programme remains mostly cloaked in secrecy, but since 2016 it has been testing drones near Cambridge, even claiming to have completed the world's first Amazon drone delivery to a 'real' customer. In the US, it has now received FAA approval to fly drones, and in the UK it has been lobbying the government to set out aviation rules for autonomous drones, and has published its own proposal for how airspace should be managed.
But can drone delivery ever work for a dense city like London, which isn't known for its landing pads?
One terrestrial alternative being developed by Greenwich-based Starship Technologies is an autonomous robot drone that drives along pavements and is already being used in some areas to deliver Just Eat takeaways. The robot is a crate on wheels that's covered in cameras and sensors, with a unique solution for navigating unexpected obstacles. If the drone's autonomous navigation can't figure it out, it connects to a human operator, who can then manually remotely pilot it around the hazard.
Sign up for Which?
- The UK's largest computing and technology title, published six times a year.
- Easy, jargon-free advice so you can make the most of your tablet, laptop or smartphone.
- One-to-one support from our friendly Tech Support team, ready to respond to unlimited member queries.
Computing here, or contact our helpful customer service team today on 029 2267 0000.