Dissertation: Monitoring driver’s emotional state and attentiveness based on physiological signals can help reduce the probability of traffic crashes
Driver distraction has been reported as the major contributor of car collisions making up about 27% of all serious injuries. The probability of traffic crashes can increase by about 10 times when drivers experience strong emotions such as sadness, happiness, and anger while driving.
Xin Zuo’s dissertation focuses on the automatic recognition of the driver's emotional states and attentiveness based on physiological signals.
“Since driving is an activity of driver-vehicle-environment interaction, drivers play an important role in this process. If there are changes in the driver’s emotional state or attentiveness while driving, it affects not only the decision making but also the driver’s actions on the driving environment.”, Zuo says.
The automatic recognition of the driver's states while driving allows the driver to make the appropriate adjustments, thereby reducing the potential for car accidents.
In her dissertation Zuo studies the fluctuation patterns of physiological signals with multiscale entropy of the optimized sample rate under different mental states to detect the driver’s abnormal state with long short-term memory (LSTM) network.
What this means is that the changes in our brains, hearts, and muscles related to our mental status, like paying attention, sad, and happy, are observed. The changes are then used to distinguish when we are for example sad or happy and when we lose our attention.
The proposed framework could achieve promising performance in analyzing the physiological changes in different driving states and detecting driver’s abnormal state automatically with different datasets, which indicates the potential of understanding and detecting driver’s abnormal state with multiple signals in practical application.
Under test conditions in this thesis, devices like electrode cap and chest Lead III are attached to subjects. on this study, a Bluetooth headband or a bracelet linked to the car's software could be developed to assist the driver.
“Such a device would help monitor the driver's physiological state and emotional states in real time, allowing the car to alert the driver to pay attention to driving if changes in his or her state are detected. For example, the vehicle's autopilot mode could be activated automatically if the driver is deemed to be in an abnormal state.”, Zuo explains.
Although visual data from cameras is already in use in many vehicles today to monitor the driver’s attentiveness and fatigue. According to Zuo, however, using physiological data is more accurate because it reflects the real internal state of human body without time delay and can't be hidden on purpose.
Xin Zuo defends her doctoral dissertation “Automatic detection of driver's abnormal state based on physiological signals”. Opponent is Professor Jari Hyttinen (University of Tampere) and custos is Professor Timo Hämäläinen (Ģֱ).
The language of the dissertation is English. or in the lecture hall.
