Electroencephalogram multi-modal fusion method for human-machine co-driving experience optimization of intelligent driving system

Abstract

The application discloses a brain electrical multi-modal fusion method for human-machine co-driving experience optimization of an intelligent driving system, and specifically comprises the following steps: synchronously collecting brain electrical signals of a driver and vehicle intelligent driving system state data; pre-processing the brain electrical signals and decoding multi-dimensional psychological state indexes; automatically identifying key interaction scenes from the vehicle state data; time-aligning the psychological state indexes and the vehicle state data corresponding to the scenes, and analyzing the causal relationship between the psychological state fluctuation and vehicle control parameters in a causal inference manner to identify key parameters and sensitive boundaries causing abnormal psychological fluctuation; and inputting the identified key parameters and sensitive boundaries into a virtual simulation environment to simulate psychological state responses under different parameter combinations and generate optimal control parameter suggestions. The application realizes a leap from responsive optimization to predictive optimization, and provides an experience optimization measure that can be previewed and verified for the intelligent driving system.

Images