How to keep drivers engaged while supervising driving automation? A literature survey and categorisation of six solution areas

Driving Activity Recognition Using UWB Radar and Deep Neural Networks

In-car activity monitoring is a key enabler of various automotive safety functions. Existing approaches are largely based on vision systems. Radar, however, can provide a low-cost, privacy-preserving alternative. To this day, such systems based on the radar are not widely researched. In our work, we introduce a novel approach that uses the Doppler signal of an ultra-wideband (UWB) radar as an input to deep neural networks for the classification of driving activities. In contrast to previous work in the domain, we focus on generalization to unseen persons and make a new radar driving activity dataset (RaDA) available to the scientific community to encourage comparison and the benchmarking of future methods.

The Origins of Passive, Active, and Sleep-Related Fatigue

Driving is a safety-critical task that requires an alert and vigilant driver. Most research on the topic of vigilance has focused on its proximate causes, namely low arousal and resource expenditure. The present article aims to build upon previous work by discussing the ultimate causes, or the processes that tend to precede low arousal and resource expenditure. The authors review different aspects of fatigue that contribute to a loss of vigilance and how they tend to occur; specifically, the neurochemistry of passive fatigue, the electrophysiology of active fatigue, and the chronobiology of sleep-related fatigue.

Effects of gender, age, experience, and practice on driver reaction and acceptance of traffic jam chauffeur systems

This study conducted a driving simulation experiment to compare four automated driving systems (ADS) designs during lane change demanding traffic situations on highways while accounting for the drivers’ gender, age, experience, and practice. A lane-change maneuver was required when the automated vehicle approaches traffic congestion on the left-hand lane. ADS-1 can only reduce the speed to synchronize with the congestion. ADS-2 reduces the speed and issues an optional request to intervene, advising the driver to change lanes manually. ADS-3 offers to overtake the congestion autonomously if the driver approves it. ADS-4 overtakes the congestion autonomously without the driver’s approval. Results of drivers’ reaction, acceptance, and trust indicated that differences between ADS designs increase when considering the combined effect of drivers’ demographic factors more than the individual effect of each factor. However, the more ADS seems to have driver-like capacities, the more impact of demographic factors is expected. While preliminary, these findings may help us understand how ADS users’ behavior can differ based on the interaction between human demographic factors and system design.

Exploration of Driver Posture Monitoring Using Pressure Sensors with Lower Resolution

Pressure sensors are good candidates for measuring driver postural information, which is indicative for identifying driver’s intention and seating posture. However, monitoring systems based on pressure sensors must overcome the price barriers in order to be practically feasible. This study, therefore, was dedicated to explore the possibility of using pressure sensors with lower resolution for driver posture monitoring. We proposed pressure features including center of pressure, contact area proportion, and pressure ratios to recognize five typical trunk postures, two typical left foot postures, and three typical right foot postures. The features from lower-resolution mapping were compared with those from high-resolution Xsensor pressure mats on the backrest and seat pan. We applied five different supervised machine-learning techniques to recognize the postures of each body part and used leave-one-out cross-validation to evaluate their performance. A uniform sampling method was used to reduce number of pressure sensors, and five new layouts were tested by using the best classifier. Results showed that the random forest classifier outperformed the other classifiers with an average classification accuracy of 86% using the original pressure mats and 85% when only 8% of the pressure sensors were available. This study demonstrates the feasibility of using fewer pressure sensors for driver posture monitoring and suggests research directions for better sensor designs.

Dynamic scan paths investigations under manual and highly automated driving

Active visual scanning of the scene is a key task-element in all forms of human locomotion. In the field of driving, steering (lateral control) and speed adjustments (longitudinal control) models are largely based on drivers’ visual inputs. Despite knowledge gained on gaze behaviour behind the wheel, our understanding of the sequential aspects of the gaze strategies that actively sample that input remains restricted. Here, we apply scan path analysis to investigate sequences of visual scanning in manual and highly automated simulated driving. Five stereotypical visual sequences were identified under manual driving: forward polling (i.e. far road explorations), guidance, backwards polling (i.e. near road explorations), scenery and speed monitoring scan paths. Previously undocumented backwards polling scan paths were the most frequent. Under highly automated driving backwards polling scan paths relative frequency decreased, guidance scan paths relative frequency increased, and automation supervision specific scan paths appeared. The results shed new light on the gaze patterns engaged while driving. Methodological and empirical questions for future studies are discussed.

Drivers use active gaze to monitor waypoints during automated driving

Automated vehicles (AVs) will change the role of the driver, from actively controlling the vehicle to primarily monitoring it. Removing the driver from the control loop could fundamentally change the way that drivers sample visual information from the scene, and in particular, alter the gaze patterns generated when under AV control. To better understand how automation affects gaze patterns this experiment used tightly controlled experimental conditions with a series of transitions from 'Manual' control to 'Automated' vehicle control. Automated trials were produced using either a 'Replay' of the driver's own steering trajectories or standard 'Stock' trials that were identical for all participants. Gaze patterns produced during Manual and Automated conditions were recorded and compared. Overall the gaze patterns across conditions were very similar, but detailed analysis shows that drivers looked slightly further ahead (increased gaze time headway) during Automation with only small differences between Stock and Replay trials. A novel mixture modelling method decomposed gaze patterns into two distinct categories and revealed that the gaze time headway increased during Automation. Further analyses revealed that while there was a general shift to look further ahead (and fixate the bend entry earlier) when under automated vehicle control, similar waypoint-tracking gaze patterns were produced during Manual driving and Automation. The consistency of gaze patterns across driving modes suggests that active-gaze models (developed for manual driving) might be useful for monitoring driver engagement during Automated driving, with deviations in gaze behaviour from what would be expected during manual control potentially indicating that a driver is not closely monitoring the automated system.