While I’m only thinking of letting a computer push the brakes of a vehicle, some are thinking more . Google is experimenting with cars that drive them selves. Until now they were always manned with a driver who can correct if the car is about to do something stupid. Some challenges have been helt with unmanned cars, but they were not on the road.
One of the benefits of the autonomous car is said to be a more efficient utilization of the road . Most efficiency increase will be created only if all cars are autonomous. Take for example the situation given in  with all the cars at a traffic light starting at the same time, instead of one after the other. In this view, a non autonomous car will be expelled from traffic. One day or the other, people who don’t feel safe in a driverless car will have no other choice.
Is this an argument against the driverless cars?
So I will build a system that can detect a human in a truck driver’s dead spot. But what should be done with the information that is sent out by this system. The first thing one can think of, is sounding an alarm. But the driver may be used to an alarm that is already sounding way to much. Also for an audible alarm it may be not immediatly clear where it is coming from. Is a truck driver – or actually any human – capable of processing all the signals a driver has to process? If a human is spotted in the dead spot and the truck detects it is steered towards that human, it is technically possible to automatically brake, thus possibly preventing a deadly accident. But does society want trucks, or cars, with a mind of their own? There may be a future in which people look back in fear to the time when trucks where driven by humans.
My first post will discuss what the goals of my thesis are. Besides my research in hard- and software, which I will not discuss on this blog, I will use the things I will study to create a device which will alarm a truck driver if a person is in it’s blind spot. The driver should only be alarmed if the object in the blind spot is a human, and not if it is a pole, tree or a curbstone. This will be done with a camera with image-recognition software.