Formerly in personal cars with safety systems we could count on mainly ABS and then also on brake assist. Modern models are already much more advanced in this respect, and have sensors, as well as cameras, including infrared sensors that detect objects or pedestrians on the road and warn the driver. However, they are not limited to warnings and if they do not react, they will undertake specific actions themselves.
The biggest problem of these systems, however, is the detection of pedestrians, as it currently takes quite a long time, which in the case of podbramkowych, when the reaction has to take place immediately, does not pass the exam. But researchers from the University of San Diego have developed a completely new algorithm that greatly improves security.
The key to efficient operation of the system is to reduce the field of vision and remove from it areas where the foot is not likely to appear, ie the roadside or sky. In addition, in the last phase of the operation, a machine learning technology is used that analyzes the image, recognizing the objects on it.