Friday, 22 September 2017

Are Driverless Cars Our Best Bet For Safer Roads?

It is believed that the ultimate road safety feature that we will see is that of automation, given that Some ninety percent of motor vehicle crashes are caused at least in part by human error. Automation of the driving task is unquestionably one of the most important issues occupying the mind of vehicle developers and road authorities and many organisations in adjacent sectors. All are keen to capture the promised value in improved safety, cleaner travel, greater efficiency, and to tackle real societal challenges posed by a growing, ageing and increasingly urbanised population. Technology is causing boundaries to be blurred. Where once there was a common understanding of the terms ‘car’, ‘road’ and ‘driver’, new categories of vehicle are emerging that challenge these previously fixed concepts.

However, the race to deliver effective automated vehicle mobility services is global and highly competitive. There have certainly been some impressive demonstrations including Otto (owned by Uber) transporting 50,000 beers in a truck that completed 120 highway miles with no human intervention. Google, now operating through its recently launched mobility services brand, Waymo, has just surpassed three million miles of automated operation across a fleet of vehicles in various locations across the US. Tesla is collecting data from drivers of its vehicles equipped with the AutoPilot partial automation functionality, giving access to data from over a billion miles of vehicles equipped with automation systems.

This new piece of tech has promised us Safer and greener driving. At least The lab coats from the automotive industry certainly appear to think so. Their confidence is based on one founding premise: that smart technologies can operate cars a whole lot better and more efficiently than people.

Imagine a car that can communicate with the cloud to identify the location of accidents or road congestion ahead, and then automatically re-route, for instance. Or put yourself in a vehicle that can "talk" to traffic lights wirelessly and regulate your speed so as to hit a green light every time. That's very efficient because when you're stopping and starting that's when you have the most load on the engine, which means more fuel use.

It's the promise of greater safety where autonomous driving really comes into its own, industry advocates claim. Ultimately, most accidents happen because of human error. Computers don't get sleepy or distracted, or take their eyes off the road because they want to change the radio station or make a phone call. Note that Google's self-drive vehicle has never so much as nudged another car.

These achievements are truly as exciting as they are remarkable, and human casualties involving road accidents will certainly fall as a direct result of this. However, the way I see it, the industries can further make automated vehicles more safe in two ways. Firstly, they must be able to articulate clearly the relative safety of automated vehicles when compared to the behavior of human drivers. This is a far from trivial task but being able to illustrate how automated vehicles avoid many of the common situations that result in road collisions will be of paramount importance in reinforcing the value of pursuing their development and use. Secondly, companies involved in the development of automated vehicles should be prepared to share safety-relevant vehicle data with industry competitors. Of course, it could be that people just like being in control behind the wheel. Autonomy, after all, is a very personal thing.

Pratyush IIT Delhi
IIT Delhi