close
MENU
Hot Topic EARNINGS
Hot Topic EARNINGS
4 mins to read

Driver-less cars


MOTORING The General Motors EN-V is an electric vehicle that owes a lot to the Segway. PLUS: Google's self-driving car (with VIDEOS).

 

David Linklater
Fri, 05 Oct 2012

MOTORING

The General Motors EN-V is the type of vehicle which many environmentally conscious people would like us to be driving in the next decade.

Or not driving as the case may be, because one of the most significant things about EN-V is that it’s capable of acting as an autonomous vehicle: it can drive itself from A to B, safely through traffic, to a predetermined destination.

EN-V is an electric vehicle that owes a lot to the Segway, because it balances on just two wheels. In fact, it owes a lot to a joint-venture concept vehicle created a few years ago by GM and Segway, called Puma.

The EN-V is small, clean and suitably futuristic looking. But its autonomous abilities are not just concept-car theory: it really can drive itself.

The idea is that a user can instruct the vehicle to go away and park itself once a destination has been reached and then retrieve it again using a smartphone application.

It’s like a plug-in pet: it will come straight away when you call.

A fleet of EN-Vs can also travel in convoy, with vehicles behind linking wirelessly with the one in front and mimicking its movements – while still using on-board sensors to avoid obstacles and pay attention to pedestrians/traffic.

More a personal mobility device 
Granted, it might be a stretch to call EN-V a car. It’s more in the line of a personal mobility device.

But the technology that allows it to drive autonomously was developed from a much more conventional vehicle: a Chevrolet Tahoe SUV adapted by GM and Carnegie Mellon University (CMU) to compete in the 2007 DARPA Challenge.

DARPA (Defense Advanced Research Projects Agency) holds an annual competition for driverless vehicles, with a view to discovering useful technology for military ground vehicles.

The 2007 event was held around a mock urban course of 96km, which had to be completed in six hours. It involved complex navigation, obeying traffic laws and negotiating obstacles.

The GM/CMU team enhanced the specified course with extra waypoints to help their vehicle and ultimately claimed victory in the event.

The Urban Challenge, as the 2007 DARPA event became known, proved that autonomous vehicles could potentially work in real-world, ever-changing environments.

Computer controlled for 500,000km
Google has taken that concept almost to the limit, running a fleet of 12 autonomous vehicles that have covered almost 500,000km solely under computer control, without a single accident.

The cars use detailed maps supplied by Google Street View, radar, video cameras and laser range finders to negotiate their way through traffic. The vehicles don’t actually travel sans occupants.

At least two people are always on board for safety’s sake, although the overwhelming success of the project thus far has led the company to consider letting the cars loose with just one occupant.

An autonomous car has to know what other vehicles are doing to operate safely and efficiently. Ultimately, the key to self-drive vehicles will be getting them to communicate with each other.

BMW, for example, is already working on a protocol called Car-To-X which will allow electronic networking (through cellphones or wi-fi) of different vehicles and roadside infrastructure.

Has application for driverless cars
Its most obvious purpose might be more efficient traffic flow – a free parking space around the corner might identify itself to you, for example – but clearly, the technology has application for driverless cars.

In Ann Arbor, Michigan, the US Department of Transport recently launched a pilot project where communication equipment was fitted to 3000 cars, trucks and buses so the vehicles could talk to each other.

The equipment exchanges information about traffic speed, location and infrastructure such as traffic lights. Loudspeakers in each car warn of potential danger or rapidly changing traffic conditions.

Cars helping each other is a nice thought. Cars relying on each other – rather than the driver – is a big step. But it’s one Volvo has taken with a project called Safe Road Trains for the Environment (Sartre).

Building on existing safety systems in Volvo cars such as active cruise control and collision warning, Sartre has developed technology that allows a convoy of cars to automatically follow a lead vehicle that is being driven manually. Just like an EN-V fleet – but so much faster.

The project completed extensive testing on public roads in Spain, with a road train consisting of a truck and three Volvo vehicles. The automated cars communicate with the lead vehicle and mimic its behaviour.

The train has covered up to 200km in a single journey, with cars following as close as five metres. Volvo says the aim of Sartre is enhanced safety but also a future where occupants can do other things instead of actually driving the vehicle.

With advances in car-to-car communication and a few changes to roading infrastructure, that day may not be far away.

David Linklater
Fri, 05 Oct 2012
© All content copyright NBR. Do not reproduce in any form without permission, even if you have a paid subscription.
Driver-less cars
24401
false