Motors Future/driverless cars:In the second in a series on the future of the car, Michael McAleer, Motoring editor, tries out a driverless car and suggests it's closer to reality than you'd think
Can you imagine a world where cars never crashed? Where you can go for a few drinks after work, and still take the car home? It's hard to foresee, particularly if you drive in Ireland.
Given the varied standard of driving on our roads, it's hardly surprising that some dream of a day when everyone in the car becomes a passenger and the driving is left to computers.
It may seem to be up there with the flying car in terms of sci-fi fantasy, but such vehicles exist today and work is being done to make them a reality within the next 15 years.
At Fiat's research and development centre in Turin, a fleet of Pandas scurry around a collection of internal roads, the steering wheel spinning and pedals moving, but there is no one sitting in the driving seat.
The first time you get into the passenger seat, even the most confident technology pioneer can be forgiven for feeling a little unnerved by the whole effect. Man must give way to machine. And that's nothing to the time they ask you to step out in front of one of these cars at a zebra crossing installed to demonstrate the power of the car's radar sensors.
It's only afterwards that they let you know of the radar sensor limitations.
Fiat's aspiration is to create fleets of urban transport solutions, with cars that return to a central location automatically or buses that run on fixed routes. However, the driving force - if you'll excuse the pun - behind driverless technology is, of course, the military. Armies lose soldiers every year in crashes and shunts, not to mind dangerous supply convoy missions.
Take the driver from the convoy into western Afghanistan, for example, and you could potentially save thousands of lives.
That's why the US government's Defense Advanced Research Projects Agency (Darpa) is so actively involved. The agency, which previously worked on the initial concept of the internet as a means of ensuring communications in the event of a nuclear attack, is now sponsoring work by researchers and car firms in the creation of vehicles that can find their own way between locations without remote control - or running into people or buildings along the way.
Enter the robot engineers. These specialists not only work on creating smart toys that mimic human behaviour, but also work on the principle of removing human involvement in jobs where the "three Ds" are at play. These stand for: dirty, difficult and dangerous.
While the average drive to work may not be dirty and most of us regard it as not that difficult, there's no doubt that given the annual road death statistics, there's certainly an element of danger.
Since 2002, Darpa has been inviting companies and senior colleges to come together with a working driverless car for a competition in California. Car companies such as Volkswagen, Toyota and Land Rover have joined elite US colleges like MIT, Carnegie Mellon and Stanford to create a fully automated vehicle that will find its way to a set location, and avoid all the potential problems placed in its way.
The challenges have varied from cross-desert routes to the most recent competition, a 96km urban area course created in a disused air force base in California.
The rules required the vehicles to obey all traffic regulations while negotiating with other traffic - both fellow competitors and some manned vehicles were used to provide traffic distraction. The challenge was to complete the course in less than six hours.
"We were given an overall map of the area the day before, so we could load it into the car," explains Chris Urmson, director of technology for the Tartan Racing team, winner of this year's event. "Then five minutes before the race began we were given the locations the car had to visit."
Tartan Racing was a joint effort between Carnegie Mellon, where Urmson works for the robotic engineering department, and General Motors.
With a timeframe of 15 months they took a regular Chevrolet Tahoe SUV - which they nicknamed "the Boss" - and added nearly 60 different sensors, including a 360-degree roof-mounted scanner.
While the pedals and steering are adjusted by mechanical arms, the entire operation is run from a processing unit in the back, with the capacity of 10 laptop computers.
"Boss works off regular GPS, similar to what you get in your Sat-Nav devices and then matches that with information it's getting via the sensors."
With a six-hour time limit to complete the course, Boss made it to all the locations and across the finish line in four hours and 10 minutes. Next up was Stanford University with a VW Passat estate, followed by a Ford, a Land Rover Discovery and a Toyota Prius.
The competition involved cars not only finding the best way to a particular location, and dealing with fixed traffic features like lights and pedestrian crossings, but also overtaking other vehicles. As this is still very early days in the development of crash-free cars, there was a crash between two driverless cars at one stage. "It was a bit surreal watching two computer-driven cars crash into each other," says Urmson.
THAT TOUCHES ON THE DILEMMAS facing society as the engineers come up with this new technology. How will society deal with such technology?
For a start, does it mean that owners bear little or no responsibility for any incident involving the cars? Will the car companies be liable for every infraction?
Even "Boss" has a large green button in the cabin that can be hit by occupants to override the computer and allow a human driver take control. There is no doubt that "auto-pilot" features will include such options, particularly in areas where GPS mapping is not so precise. In these instances the car could request the human driver take over.
Aside from military use, initially it's thought that this technology could be used in heavy traffic. Drivers could switch on the automatic drive feature and sit back, take calls, do some work, or surf the internet while the car crawls along, obeying traffic signals and keeping a safe distance from the cars in front. Then when the traffic lightens up, the human takes over again.
The secret to this technology is good GPS mapping, explains Urmson. Improve this, and the technology is no longer a science project.
However, it does raise another issue for car firms and society in general. If the cars know the speed limits in certain areas and the human driver exceeds it, to what level is the car firm responsible for allowing the vehicle to break the law, and to what degree should they be responsible for informing the police?
Cars today feature several corrective technologies, to rescue drivers from their own mistakes. The next step could be to remove the human element, the fallible being whose mistakes are the reason for the vast majority of fatal crashes.
The big fear for many will be putting your faith in the machine. Yet we do it regularly in a metal tube some 30,000 feet in the air, when the pilot opts for autopilot. These days, even landings can be completed without the human crew ever taking the controls.
We already have features such as stability control that steps in to correct our mistakes, adaptive cruise control that uses radar to maintain the distance between the car and the vehicle in front, self-parking cars like the new Lexus LS, and lane warning systems that read the road markings and tell us when we're crossing the white line, as in Citroën's range of cars.
Engineers are pushing the boundaries of what's possible, challenging the traditional driving techniques established over a century of motoring.
The question is whether society is ready to accept the changes, and how it responds.