Driver Fatigue Detection Based on SVM and Steering Wheel Angle Characteristics

In order to overcome the current driving fatigue method and system limitations, the paper proposes a driving fatigue detection method which is based on steering wheel angle features and support vector machines. This method evaluates the driving fatigue based on steering wheel angle, using AR model to analyze the time sequence and extract parameters from fixed-order model as the input vector, dividing the driving fatigue state into three degrees as output vector by the Stanford sleepiness scale, then using SVM as the classifier model, and optimizing SVM parameters with the method of CV on the basis of supporting vector machine theory. Finally, verified by specific examples, this method can effectively detect driving fatigue with a higher detection rate, so it has certain engineering application value.