As automobiles become increasingly autonomous, there is growing interest in personalized automation based on human behavior. However, there is an increasing need for behavioral data from vehicle occupants that could be acquired through physiological sensors. Hence, the motivation of this review is to provide information to develop practical physiological sensors to assess behavior, while maintaining the acceptance, privacy, and safety of vehicle occupants. In order to improve driver and passenger experiences and build trust between autonomous vehicles and humans, the current review considers a number of studies addressing driving tasks based on sensors that monitor vehicle occupant behavior. By including applications to emergency and non-emergency driving scenarios, a survey of literature is conducted on physiological sensing for personalization of semi-autonomous driving. Given the limited availability of sensor technology that is specifically designed to monitor behavior in automobiles, the survey is provided as a resource to further physiological sensor development for automotive applications.

Due to complex and uncertain driver behaviors during pedestrian collision avoidance (PCA) maneuvers, it is challenging to accurately describe human driver behaviors for manual takeover of automated vehicles. In this context, this paper proposes a novel modeling method integrating neuromuscular dynamics with a long short-term memory (LSTM) network. A neuromuscular dynamics model is responsible for capturing the dynamic interaction between the steering system and the neuromuscular system of the driver, whereas a LSTM network is used to describe the visual behaviors of the driver accurately during manual takeover. A series of high-fidelity driving simulator experiments with 12 participants were conducted to acquire driver behaviors during manual takeover under PCA conditions. Scenarios with manual takeover steering and conventional manual steering were compared to analyze driver behaviors under PCA maneuvers. In terms of root mean square error, mean absolute error, and determination coefficient, the experimental results indicated that the proposed method had higher identification accuracy than a method combining two-point visual modeling and neuromuscular dynamics. The proposed method also captured the dynamic interaction between the vehicular steering system and the neuromuscular system. Thus, in the interest of developing manual takeover control systems of automated vehicles, the proposed method can better represent lateral behaviors of drivers.

Driver distraction is a well-known cause for traffic collisions worldwide. Studies have indicated that shared steering control, which actively provides haptic guidance torque on the steering wheel, effectively improves the performance of distracted drivers. Recently, adaptive shared steering control based on the forearm muscle activity of the driver has been developed, although its effect on distracted driver behavior remains unclear. To this end, a high-fidelity driving simulator experiment was conducted involving 18 participants performing double lane change tasks. The experimental conditions comprised two driver states: attentive and distracted. Under each condition, evaluations were performed on three types of haptic guidance: none (manual), fixed authority, and adaptive authority based on feedback from the forearm surface electromyography of the driver. Evaluation results indicated that, for both attentive and distracted drivers, haptic guidance with adaptive authority yielded lower driver workload and reduced lane departure risk than manual driving and fixed authority. Moreover, there was a tendency for distracted drivers to reduce grip strength on the steering wheel to follow the haptic guidance with fixed authority, resulting in a relatively shorter double lane change duration.

Contemporary advanced driver assistance system (ADAS) features for semi-autonomous vehicles include braking assistance during collision avoidance. Although pre-collision detection typically relies on sensing systems to enable production vehicles to perceive oncoming road obstacles, the physiological state of the driver is not measured to predict emergency braking. On the other hand, previous driving simulation experiments have demonstrated the ability of regularized linear discriminant analysis (RLDA) to predict pre-collision braking using brain signals from multiple electroencephalogram (EEG) electrodes. In contrast, the current study used EEG data from these previous experiments to determine the quality of support vector machine (SVM) predictions as a first step towards realizing a brain-computer interface for emergency braking. Power spectral density (PSD) features were extracted from the EEG of one electrode to train and evaluate an SVM. Through a novel data ablation analysis, the optimal number of PSD components was determined to optimize model classification quality measured by the area under the curve (AUC). A comparison of the proposed model to the previous RLDA and other machine learning methods indicated that the SVM had a superior AU C. Thus, the proposed model is a candidate for assisting ADASs with pre-collision detection. Moreover, since the proposed model only utilized one electrode, our study potentially contributes to the facilitation of brain-computer interfaces for autonomous vehicles.