PARIS – The move toward autonomous vehicles is changing the relationship between man and machine. Man has been the master, but as cars become autonomous, man will become part of the machine’s system.

“We are going to be obliged to take into consideration the driver,” says Patrick Sega, Valeo’s director-transversal projects, research and development. “There are tasks that are going to be delegated to the vehicle, and the vehicle is going to exchange information in a more significant way with the driver. We are going to have to develop the human-machine interfaces adapted to the requirements.”

Today, a vehicle monitors itself. For example, it chimes to remind the driver the fuel level is low or when a key has been left in the ignition. In the future, when drivers allow the car to drive itself part of the time, the car has to monitor the driver to make sure he is able to take over responsibility when needed.

Auto makers and suppliers have been developing advanced systems such as lane-keeping assistance and adaptive cruise control. But since Google put the spotlight on its semi-autonomous experimental car, the industry has been thinking hard about making cars more independent.

Valeo is a major supplier of advanced automotive electronics, along with Bosch, Denso, TRW and Continental. In an interview with WardsAuto, Sega says his company’s strategy is to concentrate on low-speed automation starting with parking and traffic-jam support, rather than projects such as autonomous convoy-driving on interstates.

“Many consumers we question say there is no way they will accept a car that drives itself at 150 km/h (93 mph) on the autoroute,” Sega says. “We talk about contextual automation, automation in very precise situations that are characterized by slow speeds, and from there we will go on to situations at higher speeds.

“The situations that we think will arrive first and be accepted, safe, affordable and ‛standardizable’ will be those around parking, traffic jams and emergency braking,” he adds. “They will arrive progressively in the five or 10 years to come, probably first on premium vehicles, where customers can pay more readily for the technology.”

Valeo started with Park4U in 2008, a system that determines whether a parking space is big enough for the vehicle, then takes over the steering. The driver needs to accelerate and brake and change from forward to reverse.

The next step, and the first to be autonomous, is Park4U Remote, in which a car parks itself, controlling direction, heading and speed, commanded by a smartphone. Park4U remote is illegal in Europe for now, because the Vienna Convention signed by all the European countries and many others (but not the U.S.) requires a driver in a moving car.

Subsequent projects include traffic-jam assistance, which will be a slow-speed adaptive cruise control that follows the car in front, starting and stopping automatically; and emergency braking, in which the car slams on the brakes to avoid or mitigate the force of a collision.

From there, Valeo “will go on to urban driving, where speeds are 30 to 50 km/h (19-31 mph),” Sega says. However, he predicts urban autonomous driving won’t arrive for at least a decade.

“In town, we don’t go fast, but thousands of things happen: a bicycle, a bus, a pedestrian, lights that turn red, the person ahead that turns left but didn’t use a turn signal.”

Some researchers are working on high-speed projects. PATH in California is a project for special lanes for autonomous vehicles, and Europe has had many experiments for platooning, or convoys of vehicles that could drive closer together for aerodynamic fuel gains, but be safe because vehicle-to-vehicle communications would allow them to brake in unison.

Sega says platooning might come for heavy trucks. Fleets could calculate fuel savings, and paid, trained drivers could rest up while in the platoon and put in their behind-the-wheel hours later, increasing the delivery range.

However, Sega says, Valeo does not believe individual drivers are asking for this service.

“For us, this is a very important point,” says Sega. “We are an industrial company, and our purpose is to develop products that will sell and be adopted by the largest number possible.”

Part of the dream of autonomy, he says, is that during a voyage, “we become passengers and we can do other things like read our mail, surf on the Internet, etc.”

And the consequence is drivers become part of a system.

“We also work on what needs to be done to monitor the driver,” Sega says. “With automatic driving, when we are going to change the mode from automatic to manual…the driver must reconfigure himself from being a passenger to being a driver again.

“We (believe) the human-machine interface will be obliged to manage the driver and help him move from one mode to the other.”

A camera could be used to watch the driver, he says, and sensors could monitor the steering-wheel angle and whether his hands are on the wheel.

“The idea is that if we are capable of observing the environment, we are also capable of looking over the state of health and vigilance of the driver.”

Everyone working on autonomous driving is studying this problem. If the driver is dead or drunk or asleep at the wheel when it is time to take over, a car driving autonomously must know it and go into an emergency mode, slowing down, pulling over and stopping.

In any case, cars will have to be watching the driver, but they won’t be insisting that he is constantly alert to the road. Software will have to be intelligent, to compare different inputs and not require that a driver always is aware of the driving situation, as that would defeat the purpose of autonomy.

“We have to propose things that are reasonable to drivers,” Sega says. “A car driver is not an airplane pilot. It is Mr. and Mrs. Everybody, from 20 years old to 80 years old. We can’t ask just anything of people.”