The application relates to a multi-
modal driving
distraction information extraction method and device, which is based on multi-
modal physiological data, configures physiological feature sequences of each
modal physiological data, constructs a
distraction evolution form model by using the physiological feature sequences and nonlinear dynamics, configures constraint conditions of the
distraction evolution form model, runs the distraction evolution form model based on distraction evolution form domain
information data, and determines an initial time, a termination time and a time envelope interval of driving distraction
information extraction of a driving distraction evolution state through preset
statistical significance checking conditions of multi-modal physiological features, so that a complete distraction cycle of the distraction evolution state is captured, the dissociation dynamics and the multi-modal physiological features are perfectly fused, the driving distraction information extracted by using the time envelope interval is completely matched with an actual physiological state of distraction, accurate extraction of the distraction information is realized, and the technology has a significant influence on the field of intelligent driving and is of great significance.