IBM registered a patent for a system for transferring control of the vehicle between the driver and the autonomous car - according to the risk on the way

IBM has registered a patent for a machine learning system capable of dynamically transferring control of an autonomous car, between a human driver and a control processor integrated in the car itself, in any case of potential danger - while introducing safety measures capable of helping prevent accidents.

IBM researchers developed the system, which they patented, based on the existing understanding of how the process of cognition and response is carried out in the biological environment of the human brain. The background that these researchers brought, in the field of computational intelligence, led to the construction of a cognitive processing model and techniques that combine sensors and artificial intelligence, in order to dynamically determine the potential safety problems facing the car - and to decide whether to let the car manage itself, or transfer control to the driver.

Thus, for example, when an autonomous car experiences an operational problem, such as an improper braking system, poor visibility conditions or disruptions on the road, IBM's system may consider whether to leave driving in the hands of the driver - or automatically transfer it to self-control, due to the autonomous system's ability to handle better In the exceptional situation (US Patent #9,566,986: Controlling driving modes of self-driving vehicles).

Dr. Assaf Adi, senior manager in the field of IOT and wearable computing at IBM's research laboratory in Haifa said that "the communication between the driver and the vehicle is two-way, so if the vehicle notices that the driver is a little tired or that the expected route in the next hour is monotonous, the vehicle will suggest to the driver to Autonomous or semi-autonomous systems (such as cruise control) and change the driving environment (changing music, temperature, lighting) in order to make it easier for the driver and create a safer driving environment." Adi also added that: "The same technologies developed here in Haifa identify the driver's condition and constitute a technological infrastructure for collaborations between IBM and Olli, the i autonomous vehicle project and BMW, while creating a more intuitive and personal driving environment."

Along with the patent for the use of machine learning in dealing with the complexity involved in the dynamic operation of the transition between human driving and autonomous driving, other inventions of the company focus on helping autonomous cars better anticipate and respond more correctly to the actions of human drivers. Thus, for example, in the patent "Automatic driver modeling for integration of human-controlled vehicles into an autonomous vehicle network" (US Patent #9,361,409: Automatic driver modeling for integration of human-controlled vehicles into an autonomous vehicle network), IBM researchers describe a learning system A machine that builds a model of human driving techniques. The invention uses a unified interface, which allows autonomous cars to communicate with each other, learning and understanding better how to communicate with human drivers, as the model becomes more aware and learns more deeply.