Negative emotion recognition using multimodal physiological signals for advanced driver assistance systems

Recent advanced driver assistance systems’ (ADASs) control cars to avoid accidents, but few of them consider driver’s comfort. To realize comfortable driving, an ADAS must sense the driver’s emotions, especially when they are negative. Since emotions are reflected in a person’s physiological signals, they are informative for sensing emotions. However, it is unclear which signals are most useful for detecting a driver’s negative emotions. To examine the usefulness of each physiological signal, we implemented an emotion classifier (negative or non-negative) using sparse logistic regression for multimodal signals. This classifier was trained using a multimodal physiological signal dataset with negative emotion labels collected, while subjects were driving a vehicle. The resulting classifier successfully classifies emotions with an area under the curve of 0.74 and identifies the physiological signals that are useful for detecting negative emotions.