Radar is a technology whose origins and first uses date back to world wars. From there it has had multiple applications and technical evolution, especially aimed at increasing the signal strength. However, work carried out by a team of Princeton scientists has transformed this technology. The result has been an AI-based radar to detect objects around the corner.
It may seem trivial, but it is an important technical achievement. And as for its applications, the researchers warn that it could have intensive use in the automotive sector. Scientists have taken the radar signals that are normally considered noise. These data, which are usually discarded, also contain information, although their imprecision does not deserve attention.
However, at Princeton, they have focused on the information contained in this noise. Through a series of analyzes, they have been able to deduce which clues contain these signals. And all this has been introduced into an algorithm, which has been trained with a limited database.
In this way, they have created an AI-based radar that can guess objects invisible to the camera. Pedestrians or bicycles that appear suddenly in front of the driver can be detected in advance.
The device sends a signal at a certain angle so that it bounces as a billiard ball would. Thus, when a pedestrian has not yet appeared behind the corner of a building, the AI-based radar already senses its presence.
A useful tool for autonomous cars
The main field of application that scientists see in the automotive industry. They claim, in fact, that their invention could be made available in the next generation of vehicles if so desired. And is that the system has some advantages at an industrial level. It is based on a well-known and well-developed technology, radar. Therefore it is not expensive. To this is added that its production is easily scalable for the same reason.
Although the researchers also go further. Instead of looking only at the automotive industry, they talk about the future of it. Such a system would be really useful for autonomous cars. It would avoid potentially serious problems and many scares, in case a pedestrian suddenly crosses.
There is a whole set of new sensors aimed at facilitating autonomous driving and empowering the connected car. This Princeton job may join them. Although the competition of software technologies will always be necessary, beyond sensors, as key players in the autonomous car.