Collision systems: sophisticated monitoring systems will be able to keep track of other vehicles and initiate evasive action if drivers take their eyes off the road.
Driving along, you notice that as the car in front turns the corner, the driver is flicking carelessly through the pages of a magazine. Rather than calling the police, however, you carry on snoozing until you arrive safely at home.
The technology to allow cars to drive themselves has largely been developed. Plans for such systems, which would combine a network of sensors with a complex on-board computer, have been shown off in futuristic presentations by several car manufacturers.
Alas for sci-fi fans, though, they are unlikely to be on the road any time soon, as people still balk at the loss of control automated systems would involve, and a few technical issues remain - such as the fact that, for these automatic driving systems to work, all cars and car owners might have to use the same technology.
However, limited versions of these systems, which would help drivers to park or to avoid collisions, are beginning to come on to the market, making the life of the driver easier and safer.
Already, automatic cruise control can judge a car's distance from other vehicles using a relatively primitive form of automation, and adjust the speed to take that into account, or return the car to the driver's control.
Adrian Garrod, a consultant engineer at Roke Manor Research in the UK, owned by Siemens, believes that the next advances to these systems to come on to the market will be parking aids and pre-crash sensing.
Parking aids could use ultrasonic waves to bounce off nearby objects, allowing the car to tell the driver where he or she should be heading, or beeping if the car gets too close to objects.
Pre-crash sensing requires devices that can recognise when a collision has become unavoidable, and which can ready the car accordingly. This might involve warning the driver and passengers, preparing the air bag to inflate, and adjusting the suspension ahead of any collision.
In Roke's view, the best technology for these sensors would be radar. The company's RadarNet programme has signed up several big car manufacturers, including BMW, DaimlerChrysler and Volvo. A network of radar sensors would be positioned around the body of the car. As these sensors do not require any line of sight, they do not have to be outside the car's body, but can be hidden inside its trim.
Cars would have four such sensors at the front, four at the rear and two on either side, all connected to an on-board computer.
At present, in prototype, the sensors are about 7.5 cubic centimetres in volume, but the final versions should have volumes of between 2 and 3 cubic centimetres.
A system of this kind would cost about £1,000 to £2,000, Mr Garrod estimates, but lower end systems with fewer features could be sold for between £100 to £200.
Simon Smith of PA Consulting Group warns of the limitations of such systems, however: "In-vehicle sensors provide a means of easing the driving and parking task and making them safer for the occupant and the vehicle. But just like a human, they can do nothing about the things they cannot see or sense. They cannot see around blind corners to detect accidents."
What if other cars are using similar technology? Is there not a danger that a car could easily confuse the signals its system of sensors is sending and receiving to its central computer with the signals of other nearby vehicles, with potentially disastrous consequences?
Mr Garrod says not, as the frequency allocation for these communications has been set by international agreement at 76 to 77 GHz.
As each of these systems would be able to use the entire spectrum allocation, he calculates that there is only a one in 10,000 chance that a car could mistake a signal from another car for one of its own, and in any case this false positive would be ignored by the car's computer as it would require a majority of its sensors to agree on a signal before counting it.
Systems like this would also be set to have a range of 200 metres at the most.
Motorola takes a different approach than Siemens, preferring to rely on digital cameras rather than radar sensors for its "Aware Vehicle" initiative.
These cameras, each about the size of a fist in prototype, would be fitted at front and rear and one to each side. They could be mounted in headlights or tail lights and door frames to avoid unsightly obtrusions.
Dean Hopkins, director of marketing at Motorola, says the company opted for visual sensors because of the amount of feedback they allow for. However, cars might also be equipped with radar or ultrasonic sensors to increase the amount of information available. Systems incorporating this technology should become available in 2008 or 2009, he says, and would be likely to be priced at between $300 and $500 per car.
Though systems with limited capabilities, like parking aids and anti-collision technology, may be on the way, most experts believe that the prospect of cars driving themselves remains very remote. "Technology for automating parts of driving has been progressing at a rapid rate for the last 15 years, but all the same, pressing a button to go from A to B is not something we see drivers wanting to do in the near future," observes Mr Hopkins of Motorola.
One serious worry is that drivers may grow too reliant on automatic systems, making them vulnerable should the technology fail. But the main resistance to systems that can partly take over the control of the car, or even assist in its control, seems to be purely psychological.
John Cheese, head of transport solutions at Detica, a UK technology consultancy, points out: "These systems have great potential safety benefits. But they fundamentally alter the relationship between the driver and the car.
"For many, the whole pleasure of driving comes from exercising freedom of control and choice over speed, for example. Research suggests that older drivers are more receptive to having technology exercise some control of the car. Younger drivers might take more convincing."
