How clever is Ford’s crash proof car?

It's another step along the road towards the driverless car. The challenges in this process were outlined at the recent Electronics Design Show Conference by Freescale's Andy Birnie. He pointed out there are a number of levels involved in automated control. At the high end are two important categories. Highly automated control is when the driver has handed over to the car, but the car is 'clever enough' to hand control back in certain circumstances. Fully automated control means the vehicle has entire control. Ford's system appears to take a different approach to this by assuming control if the driver doesn't react to the situation unfolding in front of them. It's all very well demonstrating such technology on a proving ground; it's a bit different allowing a driverless car – or one which has the potential to take control in certain circumstances – on suburban streets. If the driver of Ford's system fails to steer or brake following warnings, the system will automatically steer and brake to avoid a collision. Working out the car in front is getting close quickly is relatively easy, but what is the 'escape plan'? With potentially milliseconds to make a decision, how will the car work out where to go? And will it know that hitting the car in front might be a safer option than swerving into incoming traffic? Such intelligent vehicles will need to be equipped with an array of smart sensors. The output from those sensors will need to be processed quickly in order for the control decision to be made in a timely fashion. But underlying all this is software. And as a contributor to the conference session asked: "Who in the room has written bug free code?" And if the system does go wrong – say the car mistakenly takes control and does 'something stupid' – who will be liable? Will it be the driver? Will it be the car manufacturer? Will it be some software developer? One thing is for certain; it will be some time before we see driverless cars on the street.