From Google to Uber to Nissan, dozens of companies are working on the concept of self-driving cars with varying degrees of autonomy. All have different end goals — from reducing the number of cars on the road to making cars more efficient and safe. However, as Friday’s crash involving an Uber driverless car in Arizona — in which the other car was at fault, but which still led to Uber suspending the testing of its driverless cars — has shown, the ride is likely to be bumpy.
To begin with, what are self-driving cars?
Self-driving — or driverless or robotic — cars use artificial intelligence and computing power to navigate from point A to point B, sensing their environment and making decisions in real time to avert mishaps. They do this using a range of sensors and cameras on all sides, as well as predetermined digital maps and real time inputs about traffic and obstacles. The first experimental self-driving cars started in 1984, but commercial trials are just a few years old. Mercedes Benz, one of the first players in the space, showcased its concept car, the F015, at the Consumer Electronics Show in Las Vegas in 2015. The US is the hotbed of this technology, with at least 27 companies — including Tesla, Google and Volkswagen — seeking permissions for tests in California alone. In August 2016, Uber acquired the self-driving technology company Otto, and in October used an Otto self-driving truck to deliver a consignment of Budweiser beer in Colorado.
But why do we need driverless cars?
In theory, it is great economics. The 120 crore cars in the world sit idle for almost 95% of the time, making them perhaps the most inefficient human creation ever. Uber thinks that if all cars were autonomous and shared to scale, we could eliminate 90% of these vehicles. For residents of cities from Los Angeles to New Delhi, traumatised by traffic jams, this is an attractive idea. Fewer cars also mean reduced need for parking spaces — the US alone as over 2 billion — and a drastic reduction in pollution. Plus, it is possible that self-driving cars would eliminate the human errors that kill thousands on the roads every year.
Are there any special laws for these cars?
They are still being ironed out. Recently, California became one of the first US states to draft a set of broad rules for the “autonomous test vehicles” on its roads. Among other requirements, the company running a test vehicle must put a driver behind the wheel as a fail safe, and make it known to everyone on the road that the vehicle is an autonomous one, and whether it is running in remote mode at a given point in time.
What are manufacturers of driverless cars trying to achieve?
Different things. Uber’s vision is to take reliable transportation to everyone, everywhere and everything. Nissan wants self-driving capabilities to make life easier for the driver, but not, unlike Uber, to replace her. Mercedes-Benz wants to help consumers put their time to better use, turning the commute into an opportunity to share lounge space. Since passengers won’t need to keep an eye on the road, they can use the travel time to relax or work.
So what’s in the way of realising these?
Google’s self-driving car has been caught in multiple collisions, and the crash in Arizona involving an Uber SUV last week was, even though no one was hurt, pretty serious. The technology is clearly nascent, and subject to other factors such as errors made by humans in other cars. There is bitter competition — Otto, Uber’s startup acquisition, is being sued by Google’s Waymo self-driving car unit, which claims it stole the company’s proprietary design for its laser-based radar system. Otto co-founder Anthony Levandowski is a former Waymo employee.
EVERYONE TO ANYTHING
UBER sees its acquisition of Otto as a way to extend its vision of transporting “everyone, everywhere” to moving around “anything”. According to Matt Sweeney, who heads product at Uber’s Advanced Technology Group, “the truck and passenger vehicles share software and hardware components, and this helps us work on different aspects of the self-driving problem at the same time”. But while passenger vehicles are for urban environments where Uber focuses on ridesharing, the trucks will run on high-speed highways where the constraints are different.
OTTO trucks are powered by a rig of radars, laser scanners and stereo cameras that give the computers inside a live 360-degree vision of the vehicle’s environment with high dynamic range, high resolution and high colour information imagery. As a result, the computing task is “almost as big as a small data centre”. Sweeney says while the truck has as many cellphone radios as possible, the bandwidth limitations are still very real, as gigabytes of data come in every minute, and “a lot of the computing has to be done in the vehicle itself”.
INDIA, with its chaotic driving environment, presents a slew of challenges, accepts Sweeney. “Those are situations we think we will be able to tackle, but those are not the first places we will go,” he said.