A Portable Multi-Modal Cushion for Continuous Monitoring of a Driver's Vital Signs.

With higher levels of automation in vehicles, the need for robust driver monitoring systems increases, since it must be ensured that the driver can intervene at any moment. Drowsiness, stress and alcohol are still the main sources of driver distraction. However, physiological problems such as heart attacks and strokes also exhibit a significant risk for driver safety, especially with respect to the ageing population. In this paper, a portable cushion with four sensor units with multiple measurement modalities is presented. Capacitive electrocardiography, reflective photophlethysmography, magnetic induction measurement and seismocardiography are performed with the embedded sensors. The device can monitor the heart and respiratory rates of a vehicle driver. The promising results of the first proof-of-concept study with twenty participants in a driving simulator not only demonstrate the accuracy of the heart (above 70% of medical-grade heart rate estimations according to IEC 60601-2-27) and respiratory rate measurements (around 30% with errors below 2 BPM), but also that the cushion might be useful to monitor morphological changes in the capacitive electrocardiogram in some cases. The measurements can potentially be used to detect drowsiness and stress and thus the fitness of the driver, since heart rate variability and breathing rate variability can be captured. They are also useful for the early prediction of cardiovascular diseases, one of the main reasons for premature death. The data are publicly available in the UnoVis dataset.

Predicting atrial fibrillation episodes with rapid ventricular rates associated with low levels of activity.

BACKGROUND: Rapid and irregular ventricular rates (RVR) are an important consequence of atrial fibrillation (AF). Raw accelerometry data in combination with electrocardiogram (ECG) data have the potential to distinguish inappropriate from appropriate tachycardia in AF. This can allow for the development of a just-in-time intervention for clinical treatments of AF events. The objective of this study is to develop a machine learning algorithm that can distinguish episodes of AF with RVR that are associated with low levels of activity. METHODS: This study involves 45 patients with persistent or paroxysmal AF. The ECG and accelerometer data were recorded continuously for up to 3 weeks. The prediction of AF episodes with RVR and low activity was achieved using a deterministic probabilistic finite-state automata (DPFA)-based approach. Rapid and irregular ventricular rate (RVR) is defined as having heart rates (HR) greater than 110 beats per minute (BPM) and high activity is defined as greater than 0.75 quantile of the activity level. The AF events were annotated using the FDA-cleared BeatLogic algorithm. Various time intervals prior to the events were used to determine the longest prediction intervals for predicting AF with RVR episodes associated with low levels of activity. RESULTS: Among the 961 annotated AF events, 292 met the criterion for RVR episode. There were 176 and 116 episodes with low and high activity levels respectively. Out of the 961 AF episodes, 770 (80.1%) were used in the training data set and the remaining 191 intervals were held out for testing. The model was able to predict AF with RVR and low activity up to 4.5 min before the events. The mean prediction performance gradually decreased as the time to events increased. The overall Area under the ROC Curve (AUC) for the model lies within the range of 0.67-0.78. CONCLUSION: The DPFA algorithm can predict AF with RVR associated with low levels of activity up to 4.5 min before the onset of the event. This would enable the development of just-in-time interventions that could reduce the morbidity and mortality associated with AF and other similar arrhythmias.

Designing for the Extremes: Modeling Drivers' Response Time to Take Back Control From Automation Using Bayesian Quantile Regression.

OBJECTIVE: Understanding the factors that affect drivers' response time in takeover from automation can help guide the design of vehicle systems to aid drivers. Higher quantiles of the response time distribution might indicate a higher risk of an unsuccessful takeover. Therefore, assessments of these systems should consider upper quantiles rather than focusing on the central tendency. BACKGROUND: Drivers' responses to takeover requests can be assessed using the time it takes the driver to take over control. However, all the takeover timing studies that we could find focused on the mean response time. METHOD: A study using an advanced driving simulator evaluated the effect of takeover request timing, event type at the onset of a takeover, and visual demand on drivers' response time. A mixed effects model was fit to the data using Bayesian quantile regression. RESULTS: Takeover request timing, event type that precipitated the takeover, and the visual demand all affect driver response time. These factors affected the 85th percentile differently than the median. This was most evident in the revealed stopped vehicle event and conditions with a longer time budget and scenes with lower visual demand. CONCLUSION: Because the factors affect the quantiles of the distribution differently, a focus on the mean response can misrepresent actual system performance. The 85th percentile is an important performance metric because it reveals factors that contribute to delayed responses and potentially dangerous outcomes, and it also indicates how well the system accommodates differences between drivers.

Silent failure detection in partial automation as a function of visual attentiveness.

OBJECTIVE: Drivers using level 2 automation are able to disengage with the dynamic driving task, but must still monitor the roadway and environment and be ready to takeover on short notice. However, people are still willing to engage with non-driving related tasks, and the ways in which people manage this tradeoff are expected to vary depending on the operational design domain of the system and the nature of the task. Our aim is to model driver gaze behavior in level 2 partial driving automation when the driver is engaged in an email task on a cell phone. Both congested highway driving, traffic jams, and a hazard with a silent automation failure are considered in a driving simulator study conducted in the NADS-1 high-fidelity motion-based driving simulator. METHODS: Sequence analysis is a methodology that has grown up around social science research questions. It has developed into a powerful tool that supports intuitive visualizations, clustering analysis, covariate analyses, and Hidden Markov Models. These methods were used to create models for four different gaze behaviors and use the models to predict attention during the silent failure event. RESULTS: Predictive simulations were run with initial conditions that matched driver state just prior to the silent failure event. Actual gaze response times were observed to fall within distributions of predicted glances to the front. The three drivers with the largest glance response times were not able to take back manual control before colliding with the hazard. CONCLUSIONS: The simulated glance response time distributions can be used in more sophisticated ways when combined with other data. The glance response time probability may be conditioned on other variables like time on task, time of day, prevalence of the current behavior for this driver, or other variables. Given the flexibility of sequence analysis and the methods it supports (clustering, HMMs), future studies may benefit from its application to gaze behavior and driving performance data.

Quantifying vehicle control from physiology in type 1 diabetes.

Objective: Our goal is to measure real-world effects of at-risk driver physiology on safety-critical tasks like driving by monitoring driver behavior and physiology in real-time. Drivers with type 1 diabetes (T1D) have an elevated crash risk that is linked to abnormal blood glucose, particularly hypoglycemia. We tested the hypotheses that (1) T1D drivers would have overall impaired vehicle control behavior relative to control drivers without diabetes, (2) At-risk patterns of vehicle control in T1D drivers would be linked to at-risk, in-vehicle physiology, and (3) T1D drivers would show impaired vehicle control with more recent hypoglycemia prior to driving.Methods: Drivers (18 T1D, 14 control) were monitored continuously (4 weeks) using in-vehicle sensors (e.g., video, accelerometer, speed) and wearable continuous glucose monitors (CGMs) that measured each T1D driver's real-time blood glucose. Driver vehicle control was measured by vehicle acceleration variability (AV) across lateral (AVY, steering) and longitudinal (AVX, braking/accelerating) axes in 45-second segments (N = 61,635). Average vehicle speed for each segment was modeled as a covariate of AV and mixed-effects linear regression models were used.Results: We analyzed 3,687 drives (21,231 miles). T1D drivers had significantly higher overall AVX, Y compared to control drivers (BX = 2.5 × 10-2BY = 1.6 × 10-2, p < 0.01)-which is linked to erratic steering or swerving and harsh braking/accelerating. At-risk vehicle control patterns were particularly associated with at-risk physiology, namely hypo- and hyperglycemia (higher overall AVX,Y). Impairments from hypoglycemia persisted for hours after hypoglycemia resolved, with drivers who had hypoglycemia within 2-3 h of driving showing higher AVX and AVY. State Department of Motor Vehicle records for the 3 years preceding the study showed that at-risk T1D drivers accounted for all crashes (N = 3) and 85% of citations (N = 13) observed.Conclusions: Our results show that T1D driver risk can be linked to real-time patterns of at-risk driver physiology, particularly hypoglycemia, and driver risk can be detected during and prior to driving. Such naturalistic studies monitoring driver vehicle controls can inform methods for early detection of hypoglycemia-related driving risks, fitness to drive assessments, thereby helping to preserve safety in at-risk drivers with diabetes.

Markov Models for Detection of Ventricular Arrhythmia.

The advent of portable cardiac monitoring devices has enabled real-time analysis of cardiac signals. These devices can be used to develop algorithms for real-time detection of dangerous heart rhythms such as ventricular arrhythmias. This paper presents a Markov model based algorithm for real-time detection of ventricular tachycardia, ventricular flutter, and ventricular fibrillation episodes. The algorithm does not rely on any noise removal pre-processing or peak annotation of the original signal. When evaluated using ECG signals from three publicly available databases, the model resulted in an AUC of 0.96 and F1-score of 0.91 for 5-second long signals and an AUC of 0.97 and F1-score of 0.93 for 2-second long signals.

Driving Safety and Real-Time Glucose Monitoring in Insulin-Dependent Diabetes.

Our goal is to address the need for driver-state detection using wearable and in-vehicle sensor measurements of driver physiology and health. To address this goal, we deployed in-vehicle systems, wearable sensors, and procedures capable of quantifying real-world driving behavior and performance in at-risk drivers with insulin-dependent type 1 diabetes mellitus (DM). We applied these methodologies over 4 weeks of continuous observation to quantify differences in real-world driver behavior profiles associated with physiologic changes in drivers with DM (N=19) and without DM (N=14). Results showed that DM driver behavior changed as a function of glycemic state, particularly hypoglycemia. DM drivers often drive during at-risk physiologic states, possibly due to unawareness of impairment, which in turn may relate to blunted physiologic responses (measurable heart rate) to hypoglycemia after repeated episodes of hypoglycemia. We found that this DM driver cohort has an elevated risk of crashes and citations, which our results suggest is linked to the DM driver's own momentary physiology. Overall, our findings demonstrate a clear link between at-risk driver physiology and real-world driving. By discovering key relationships between naturalistic driving and parameters of contemporaneous physiologic changes, like glucose control, this study directly advances the goal of driver-state detection through wearable physiologic sensors as well as efforts to develop "gold standard" metrics of driver safety and an individualized approach to driver health and wellness.

A Novel Atrial Fibrillation Prediction Algorithm Applicable to Recordings from Portable Devices.

Atrial Fibrillation (AFib) is by itself a strong risk factor for many life-threatening heart diseases. An estimated 2.7 to 6.1 million people in the United States have AFib. With the aging of the U.S. population, this number is expected to increase. In this preliminary study, a heart rate-duration criteria region is proposed to automatically label symptomatic AFib events using recordings from portable ECG monitors. A Markov Chain algorithm is implemented to classify prediction intervals that are 2 minutes before the symptomatic AFib events. The method yields an overall accuracy value of 82% with 0.91 AUC.

Quantification of facial motion for objective evaluation using a high-speed three-dimensional face measurement system--a pilot study.

OBJECTIVE: To estimate asymmetry in normal and pathological facial functions using an established computer generated objective evaluation technique. METHODS: Analysis was performed on three-dimensional (3-D) data captured using a specially designed 3-D face shape measurement system. Six healthy volunteers and six patients with Bell's palsy were analyzed for forced eye closure and grinning actions. The asymmetry was computed at locations that had the greatest effect in each action on both the left and right sides of the face, in reference to a relaxed condition. The patients' data were captured and analyzed repeatedly for a period of three days for an average interval of four weeks, and the results were compared with the Yanagihara scale. RESULTS: The control set of a normal sample exhibited a low standard deviation and a high correlation coefficient in both facial actions, and this contributed to a robust evaluation. The patients showed a higher standard deviation than the healthy subjects because of the larger degree of scatter of the data points in their respective data distributions. During the three clinical examinations, our proposed quantification method produced a continuous grading scheme, as opposed to the discrete scheme of the House-Brackmann grading. CONCLUSION: Our proposed system shows advantages over the existing methods in that it does not rely on reference points nor does it use markers to analyze facial deformation. In addition, our estimations are very robust and accurate because our approach directly evaluates 3-D spatial variations in normal and pathological facial functions.