It is a future we’ve been programmed to believe. An expressway chock full of cars but no one’s driving them. We’ve seen them in Total Recall, Minority Report, and iRobot – and for some of us, they cannot come fast enough. The good news is, they’re here. But not in the scale of the above mentioned movies.
In 2014, Tesla – the electronic supercar of the future – began prepping its Model S with hardware for its self-driving feature. According to its blog, the car was equipped with ‘a forward radar, a forward-looking camera, 12 long-range ultrasonic sensors to “see” 16 feet (4.8m) around the car in all directions at all speeds, and a high-precision digitally controlled electric assist braking system’. In October 2015, Tesla announced its version 7.0 software, aka Autopilot – with “a range of new active safety and convenience features” – that could work with the hardware to deliver a safe driverless experience.
Tesla said the Model S had a “fully integrated autopilot system involving four different feedback modules: camera, radar, ultrasonics, and GPS (global positioning system). These mutually reinforcing systems offer realtime data feedback from the Tesla fleet, ensuring that the system is continually learning and improving upon itself.”
Yet despite all of these checks and balances, it did not stop Tesla’s first Autopilot fatality. In May 27 this year, Tesla enthusiast Joshua Brown was killed when the Model S he was in control of collided with a trailer. The truck driver later claimed in a Guardian report the car was going so fast, “I didn’t see him”. Chief Executive Officer Elon Musk tweeted his condolences but Tesla’s official statement sought to absolve the company of blame. Tesla had always maintained Autopilot was a beta and the responsibility rests on drivers to stay alert at all times. In fact, the car will issue visual (dashboard) and audible alerts if it does not detect the driver’s hands. The car will “gradually slow down until hands-on is detected again”, said Tesla. Scant comfort for Mr Brown’s family.
Who has right of way?
With the emergence of autonomous cars, manufacturers have to grapple with a deeper moral and ethical question: in an accident, whom should the car save? According to data collected by science website Live Science, people in the United States (US) generally want driverless cars to be utilitarian, which means it will minimise fatalities or injuries – even if it means harming the people in the car. For example, if a car with four passengers in it was heading in the direction of a stalled bus and it cannot safely engage it brakes in time, the car will make a decision to drive into a tree – or off a cliff – to save the 25 people (or so) in the bus and sacrifice its passengers.
The same study, however, noted that most of the respondents would not want to be passengers of said utilitarian cars, and do not want regulations to force utilitarianism on autonomous cars. So we’re back to square one. Build an ethical car and risk not having buyers, or protect drivers and look like a roadside bully.
However, carmakers insist autonomous cars will save more lives overall. Without human drivers, cars improve traffic efficiency (reduce traffic jams), reduce pollution (people won’t suffocate in cars in said traffic jams), and eliminate up to 90 percent of traffic accidents, according to CBS news.
So what is the solution to this problem?
All hail Skynet
The truth is self-driving cars are hardly a solo effort. It’s not just down to carmakers to revolutionise the way we commute. Cars need to ‘talk’ to other cars, and roads, and buildings, and pedestrians. In every scenario that you can think of where you might have to swerve – or jam brake – to safeguard public and private properties, as well as lives, will have to apply to self-driving cars. As Business technology website Informationweek.com said: “The best way to make roads safer is to supplement, and eventually bypass, the weakest link in every vehicle’s crash avoidance system – the easily distracted, irrational, and sometimes dangerous human driver.”
That’s where Big Data steps into the picture. When cars talk to infrastructure and people (specifically their super-tapped-in mobile phones), that’s when they can truly claim to be safe. By linking to a connected world, you eliminate variables (the biggest being human error) and make self-driving safe.
INVENT, or INter VEhicular Network Technologies, a project by the New Jersey’s Science & Technology University, has been mapping a way to use “inter-vehicular networking, computing and sensing technologies” to make self-driving cars safe since 2014.
For a transport system akin to the one in Minority Report, vehicles need to avoid each other easily. The Mobility Transformation Center at MIT (Massachusetts Institute of Technology) is looking into using radiowaves to power Vehicle-To-Vehicle (V2V) and Vehicle-To-infrastructure (V2I) to do just that. Taking their cue from nature, Engineering Professor Larry Burns said, “Bees swarm. Geese flock. And they’re not running into each other.”
(Featured photo and above via digital90210)
Since experts have been talking about Big Data and Artificial Intelligence for the last two years, why haven’t we seen more large scale trials based on that? For one, carmakers are not traditionally technology experts. And they’re not in the business of contributing to a mindless drone of transport that has no individuality. After all, cars are largely emotional purchases – and you’re more likely to buy a car that exudes your character, like a Tesla Model S.
For a transport infrastructure that swarms as in Minority Report, we may have to look at a future where carmakers sell to the government instead. And Singapore may just be the city to make that step.
According to a press release by A*Star Institute for Infocomm Research, the Land Transport Authority (LTA) is committed to research and development, as well as testing Autonomous Vehicle (AV) technology since 2014. LTA will play the regulatory role while A*Star contributes to the technology. Minister for Foreign Affairs and Minister-in-charge of the Smart Nation Initiative Dr Vivian Balakrishnan said the government is looking to trial autonomous mini-buses and vehicles, which would be available on-demand, in the next one to two years. This is to close the last-mile gap, to bring people from their homes to the Mass Rapid Transport (MRT) station or bus interchange, the minister said at IoTAsia 2016. True to his promise, Singapore set in motion the world’s first self-driving taxi in August – ahead of Uber’s trial in Pittsburg said The Straits Times.
The wealth of information gathered from Singapore’s trial and the one by Uber will pilot the way self-driving cars in the future talk to infrastructure and other vehicles. It may even pave the way for insurance companies to determine who is liable in a car accident – will it be the human driver (who’s prone to error), the smart car manufacturer or the government (who commissioned the trial)? We may even see the first iterations of Omni Consumer Products (OCP).