Globetek is a professionally managed company, which is an importer and suppliers of Current sensors, Optical comparator, Dew Point measuring instrument & Current probe in India
When will cars drive themselves?
By: Dan Keegan
Date: Friday, 05. March 2010
Attempts to educate and train drivers during the past century have been something of an embarrassment.
Sure, at one time almost every high school in North America had a “driver’s ed” program, but these programs were light on the meat of driving techniques, tactics and strategy (a few hours behind the wheel and boring hours of classroom theory). The classroom sessions dealt with everything from how to fill out the licensing forms to laws, responsibilities, rules of the road and traffic citizenship. Too much for teens mentally fixated on getting a license and achieving a new level of social independence.
In the late part of the last century and the first decade of this one the education and training of drivers became worse. It’s as if the driving powers that be gave up.
Result: road populated with drivers who poorly understand the protocols of road use and traffic, have mediocre skills, crash just as often as drivers have always crashed and, worse still, invariably consider other drivers to be the problem.
Giving up on humans
It’s often been said that the main reason driver training is so poor is that those responsible for training, and consequently for training the trainers, don’t have a very high level of knowledge or skill themselves, and that’s probably true. But always, underlying the apathy towards raising driver knowledge and skills levels, there’s been this notion that there’s only so much you can do with us intransigent and intractable humans.
Now, those who feel that way have another reason to be smug about their attitude. Cars that can drive themselves are becoming reality.
Fifty years ago, the smart brains in the driving industry were focused on cars. Drivers got little thought. Today, many very smart brains are working on the driver, but this time the driver will be a computer.
This is an idea that doesn’t sound far-fetched any more. Already, cars that drive themselves, in a limited fashion at least, have been tested. (see Driverless cars hit New York Streets). Almost every day in the media we see some news item about intelligent cars, smart roads, smart traffic systems.
So when will it all happen?
For the most part, the techologies are already there. Already we can buy cars that can parallel park themselves. Car makers are already using sensors of various kinds to feed information about obstacles, traffic lights, road surfaces, other road users, and traffic information to vehicles. Experimental ‘Intelligent highway Systems” are testing sensors and radio devices that will feed information about just about everything drivers need to know about traffic, to receivers in the vehicle.
Whether that information is fed to a human driver or a computer is a very complicated question. While the technologies are out there, and some are working away helping drivers in different ways (e.g. GPS, road sensors, traffic light sensors), putting the whole system together is an awesome task.
Very likely the first use for the computer-driven car will be convoys of vehicles using special lanes on special routes. When a vehicle joins the convoy, control will be handed over to the convoy's control computer. The driver can relax, watch a movie, use the internet, talk on the phone (see Drivers.com on driverless cars). “Much like getting on a train,” one researcher suggested, except you will be in your own vehicle.
Sensors and computers will maintain spacing between vehicles. The convoy will have its predetermined destinations. Vehicles will join and drop out of the convoy. Very likely the first commercial convoys will be on interstate highways or privately run toll roads. The idea, along with some of the experiments, was explored in a 2004 Economist article.
However, these convoys would be simple to achieve in comparison to the task of developing a completely autonomous vehicle that could take its passengers shopping, to visit friends, or to a destination in another city. For this, the vehicle’s controlling computer would have to “see” lane markings, traffic lights, pedestrians, lane obstacles, other traffic, animals — everything a human driver would see.
Human vs. Computer
Programming a computer to maneuver a vehicle (acceleration, braking, turning) is relatively easy compared with the task of programming it to a level of driving tactics and strategy that human drivers are capable of.
However, what we humans are capable of and what we actually do are two different matters. The level of ability that most of us achieve would be entirely unacceptable in a computer that could be entrusted with the same task. That means that the computer programmers will have to study driving skills, tactics and strategy at a level way beyond that considered by the authors of current driver handbooks, text books or instructor training manuals.
Undoubtedly, some very sharp minds will be focused on this task.
Groups interested in the task will be diverse - ranging from software developers to car manufacturers, highway builders, highway management groups, traffic control system manufacturers, navigation systems makers, and countless others. Even environmentalists will have a keen interest, since autonomous vehicles could, potentially, save hugely on congestion and fuel costs.
So — when?
The most optimistic guesses about when it will all happen will likely start at least 15 - 20 years into the future, although 50 years might be more realistic. “We’re in the infancy stages at this point,” TRW Automotive’s John Wilkerson told Aftermarket Business recently.
A system that can accommodate automated vehicles will be immensely complex to set up and weave together into an integrated whole. For the immediate future the most likely scenario is the gradual implementation of parts of the system - for example, sophisticated navigation and traffic information systems, automated segments such as cruise control systems that automatically maintain following distances and even brake and steer towards the best lane, traffic light warning systems, cars that communicate with one another and automatically warn of impending collisions, and systems that keep drivers alert.
Meanwhile, the need to upgrade driving techniques and the tactical abilities of drivers is greater then ever, not just because the traffic system is more complex than ever but also because the desire and the capability to multi-task while driving is now irresistible.
Further comments to this article have been disabled.
All Comments (4)
Showing 1 - 4 comments
On The Record from Hamby Hutcheson
This is intellectual crap:
“Hands-free devices offer no safety benefit when driving.” 2010 - The National Safety Council.
I was going to reject Hamby Hutchison's convoy skepticism outright until i got to the phrase
'At least not in what we currently consider a "motor vehicle". '
Like to know more about what you mean by that Hamby?
After all technology can oveercome all kinds of human obstacles but the beer party? Scary!!
When? Hopefully .... Never!
I was one of the Committee Chairs and Task Force Leaders some 20 years ago when these ideas first floated in the National Automated Highway System (NAHS) program funded by the Federal Givernment.
- "Feet Off": Here today with the current crop of cruise controls.
- "Hands Off": Here to day (to a degree) with lane change management and automated parking assist.
- "Brain Off": NEVER GOING TO HAPPEN (IN MY OPINION). At least not in what we currently consider a "motor vehicle".
Imagine the "hot dog" that gets into the "platoon", kicks back, pulls out his/her favorite mag AND a couple of six-packs and proceeds to drink until wacked and needs to now go to the potty .... with beer urgency mind you!
Exiting will be deadly ... not going to happen on my watch!
This type of "platoon" will not be on wheels and will not be limited to a pavement lane (IMO).