Applying deep neural networks for multi-level classification of driver drowsiness using Vehicle-based measures

Arefnezhad, Sadegh; Samiee, Sajjad; Eichberger, Arno; Frühwirth, Matthias; Kaufmann, Clemens; Klotz, Emma

doi:10.1016/j.eswa.2020.113778

Cited by 56 publications

(32 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We also performed comparisons with previous studies. Previous studies [27,29,31,40] used traditional machine learning-based methods, whereas Arefnezhad et al [33] proposed a CNN-LSTM network to detect drowsy driving.…”

Section: Experiments Results For the Proposed Ensemble Cnnmentioning

confidence: 99%

Section: Experiments Results For the Proposed Ensemble Cnnmentioning

confidence: 99%

“…As shown in Table 10, SWA data are used in many studies and pedal pressure data are also used to detect drowsy driving [29,31,33,40]. Although not used in this paper, in-vehicle sensor data such as lateral acceleration data [33] and vehicle speed [27] are also used. In addition, deviation of the vehicle position in a lane [27,33,40] or PERCLOS [27] are used to improve the detection performance.…”

Section: Experiments Results For the Proposed Ensemble Cnnmentioning

confidence: 99%

“…However, it is extremely dangerous to acquire sensor data from an actual vehicle during drowsy driving, and as such, previous studies have collected the data via driving simulations. In numerous studies, a limited simulation environment [33] has been used to restrict the road types or to include some special data such as lane changes and lane positions. In addition, in the collected data, the amount of drowsy driving data is low.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Ensemble CNN to Detect Drowsy Driving with In-Vehicle Sensor Data

Jeon

Kim²,

Baek

2021

Sensors

View full text Add to dashboard Cite

Drowsy driving is a major threat to the safety of drivers and road traffic. Accurate and reliable drowsy driving detection technology can reduce accidents caused by drowsy driving. In this study, we present a new method to detect drowsy driving with vehicle sensor data obtained from the steering wheel and pedal pressure. From our empirical study, we categorized drowsy driving into long-duration drowsy driving and short-duration drowsy driving. Furthermore, we propose an ensemble network model composed of convolution neural networks that can detect each type of drowsy driving. Each subnetwork is specialized to detect long- or short-duration drowsy driving using a fusion of features, obtained through time series analysis. To efficiently train the proposed network, we propose an imbalanced data-handling method that adjusts the ratio of normal driving data and drowsy driving data in the dataset by partially removing normal driving data. A dataset comprising 198.3 h of in-vehicle sensor data was acquired through a driving simulation that includes a variety of road environments such as urban environments and highways. The performance of the proposed model was evaluated with a dataset. This study achieved the detection of drowsy driving with an accuracy of up to 94.2%.

show abstract

Section: Experiments Results For the Proposed Ensemble Cnnmentioning

confidence: 99%

Section: Experiments Results For the Proposed Ensemble Cnnmentioning

confidence: 99%

Section: Experiments Results For the Proposed Ensemble Cnnmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Ensemble CNN to Detect Drowsy Driving with In-Vehicle Sensor Data

Jeon

Kim²,

Baek

2021

Sensors

View full text Add to dashboard Cite

show abstract

“…Using the detailed observation data, it was possible to determine for each time point, which posture the subject was currently adopting, which activities were performed and in what state of drowsiness the subject was at that time. These results are useful for a better understanding driving behaviour, but have also been used as training input for the development of drowsiness detection systems (Arefnezhad 2020a, Arefnezhad 2020b.…”

Section: Summary and Discussionmentioning

confidence: 99%