Autonomous driving - are we there yet? Friday, 12 August 2016

News article written by Corbett Communications. The statements made or opinions expressed do not necessarily reflect the views of Engineers Australia.

An announcement from Tesla that will be “really big news” has been flagged by its now-iconic CEO, Elon Musk, who stopped short at saying its Model 3 would be geared for Level 4 fully autonomous driving (see below for level descriptions).

The Model 3 was unveiled in March and is set to commence production late in 2017, with speculation buzzing about its level of autonomy especially now the race is on with the Obama administration promising to invest almost US$4 billion in research over the next decade.

“What we’ve got will blow people’s minds; it blows my mind,” Musk said. “It’ll come sooner than people think.”

But in the rush to not drive, will these vehicles be safe on our roads? Musk revealed that Tesla is focusing on the development of advanced, narrow AI and the improvement of advanced neural maps, but conceded there are still challenges with the software and the size of the computer needed to fit in the vehicle.

But this revelation, reported by Electrek, has come hot on the heels of the reported death of a Tesla test driver in a crash in May while the Autopilot system was engaged. The news of the first known death from an autonomous vehicle (Model S electric) has brought to light concerns about expectations around autonomous driving, putting pressure on software engineers to come up with the goods, and fast. But, then there’s the age-old question of ‘just because we can, should we?’

While the thought of driverless vehicles is exciting, “just like the movies”, the reality is that in the not-so-distant-future a lot of people will be out of a job – a known repercussion after leaps in technology since the Industrial Revolution. In the computer age, ITEE engineers’ roles are pivotal role to these changes.

Up until now, there have been a number of incidents involving self-driving cars, and not just Tesla’s. Earlier this year, a Google car drove into the side of a bus while attempting to turn off a major road and detected sand bags around a storm drain. The car manoeuvred around the obstruction, switched lanes and struck the bus. This was one of almost a dozen minor collisions around the company’s headquarters since it started testing them on the city’s streets three years ago.

A 2015 report of a Volvo XC60 autonomous vehicle running into pedestrians is shown in video from the Dominican republic where a representative from the car maker commenting later that the vehicle is not equipped with Pedestrian Detection Functionality – that’s sold extra. This poses a moral dilemma for Volvo’s software engineers involved in the design and production of these vehicles where crash avoidance with other vehicles takes precedence over hitting pedestrians.

However, the collision warning system, according to Volvo, “cannot detect all pedestrians in all situations” due to what the camera can see and the level of lighting involved. So, how safe will you be in the car or on walking down the street? Only time will tell. Meanwhile there is a disconnect between public perception meeting reality and we’re not there quite yet.

Levels of autonomous driving:
Level 1: Semi-autonomous - most functions are still controlled by a driver, but some (like braking) can be done automatically by the car.
Level 2: At this level the driver is not physically operating the car but is in the driving seat and always ready to take control.
Level 3: The vehicle is able to engage safety-critical function under certain traffic conditions, but a driver is still present but not required to monitor the situation as previous levels.
Level 4: Fully autonomous driving. The car is designed to perform all safety-critical driving functions and monitor roadway conditions for the entire trip.
Level 5: A fully autonomous vehicle without any driving controls.

Author: Desi Corbett