Woozy Driving Detection

Abstract

Abstract-wooziness has been identified as one of the primary causes of injury or death in car accidents. As a result, it is proposed in this research work to develop an adaptive heavy vehicle driver fatigue and alertness model based on spectral bands by combining signal processing algorithms with soft computing techniques such as the neuro-fuzzy algorithm to estimate the driver cognitive state while driving a vehicle in a virtual reality (vr)-based dynamic simulator under monotonous driving conditions. As a result, it is proposed in this study to use soft computing approaches to reduce the amount of features and non-linear supervised classification algorithms to classify them. The suggested adaptive model distinguishes between the driver\'s level of exhaustion by analysing brain responses to determine whether the driver is fatigued owing to task- induced variables or attitude/behaviour, and then the amount of fatigue is associated to sleepiness (i.e. Level of alertness towards driving). The adaptive model can also be used to warn drivers and regulators about impending catastrophes by optimizing the features of interface systems. The suggested system notifies the driver when he or she is fatigued or drowsy based on the identification of cognitive state and calculates the fatigue index and alertness level. The proposed method also assists the driver in being more alert and intuitive in order to avoid deadly road accidents.

Introduction

I. INTRODUCTION

Wooziness is a condition of gradually reduced consciousness linked with a desire or propensity to sleep that exists between waking and sleep. We are developing a sleepiness detecting system. A large number of people use the roadway at all hours of the day and night. Taxi drivers, bus drivers, truck drivers, and those commuting great distances are all sleep deprived. As a result, driving when drowsy is extremely risky.

Wooziness has been classified using data from a wide range of sources. Physiological parameters such as electroencephalography, heart rate variability, and respiration are examples of driver-based measures. Furthermore, eye, gaze, and head measurements have been utilized to categorize driver state. Drowsiness has also been classified via steering wheel manipulation. Distinct techniques have different advantages and disadvantages.

In vehicle- based detection systems, for example, may categorize chronic tiredness based onIndications such as unpredictable steering and lane incursions. This article describes current woozy research that was carried out using an automotive grade production-ready driver monitoring system. It starts with a description of data gathering methods as well as the methodology to developing and evaluating a wooziness algorithm.

II. RELATED WORK

Many ways have been proposed to improve the accuracy of the driver drowsiness system detection. They suggested an electroencephalography (eeg)-based approach for detecting drowsiness. For discriminating drowsiness and alertness, chaotic characteristics and the logarithm of signal energy are extracted. The classification was done using an artificial neural network, which had an accuracy of 83.3 percent. Suggested a model based on the combination of driving quality signals, eeg, and electrooculography to detect tiredness. A class separability feature selection method was utilised to find the optimal subset of characteristics.

III. OBJECTIVES

The main goal is to build a sleepiness detection system based on eye tracking, it is thought that the signs of driver weariness can be recognized this way. Early enough to prevent being involved in an automobile accident detection of exhaustion involves observing blink patterns and eye movements. The study of facial pictures is a major field of study. Face recognition and person identification are examples of uses. Systems of security the focus of this research is on the localization of the eyes, which entails inspecting the full image of the eye and a self-developed system for detecting the position of the eyes algorithm for image processing.

This proposed research intends to design a real- time, easy- to-implement, nonintrusive, and accurate drowsiness detection system to overcome the limitations of present wooziness detection approaches. The term "wooziness" refers to a condition of decreased alertness that is often accompanied by performance and psychophysiological changes and can lead to loss of awareness or falling asleep at the wheel.

IV. METHODOLOGY

A face detection method is used to recognise faces in images, which is then passed on to an eye detection algorithm as input.

Once the face has been identified, an eye identification method is utilised to extract the eye region from the facial images, which is then fed into cnn.

Deep features are extracted using a cnn with four convolutional layers, which are then sent to a fully connected layer in cnn, the soft max layer classifies the pictures as sleepy or non-sleepy.

Cnn soft maximum layer sort the photos into sleepy and non-sleepy categories.

The drowsiness detection system architecture proposed using deep cnn the proposed paradigm is divided into three Stages. Preprocessing, feature extraction,and the deep cnn classifier are the three stages of the procedure.

VI. RESEARCH GAP

In terms of accuracy and underpinnings, current research is promising; nevertheless, more study is needed to validate techniques for on-road application. Model adaptation should also be addressed, with the goal of detecting driver drowsiness early or before the driver is put at risk

VII. FUTURE WORK

In the future, we will apply transfer learning to increase the system's performance and it will be critical for studies of sleepiness detection to include participants with sleep disorders, so that they can be monitored properly on the road as well. Once these issues are resolved, accurate sleepiness tracking and, presumably, broader commercial adoption will be possible.

Conclusion

We conclude that a new strategy for detecting driver tiredness based on eye condition is proposed in this proposed work. This determines if the eye is drowsy or not, and alerts the user with an alarm if the eye is drowsy. The viola-jones detection algorithm is used to detect the face and eye region. To extract features, a stacked deep convolutional neural network is created and employed in the learning phase. The driver is classified as sleep or non- sleep using a softmax layer in a cnn classifier. The proposed system had a precision of 96.42 percent. When the model forecasts drowsy output condition continually, the proposed system 465 successfully identifies the status of the driver and alerts with an alarm.

References

[1] Jabbar, r., al-khalifa, k., kharbeche, m., alhajyaseen, w., jafari, m., & jiang, s. (2018). Real- time driver drowsiness detection for android application using deep neural networks techniques. Procedia computer science, 130, 400-407. [2] Wijnands, j. S., thompson, j., nice, K. A., aschwanden, g. D., & stevenson, m. (2020). Real- time monitoring of driver drowsiness on mobile platforms using 3d neural networks. Neural computing and applications, 32(13), 9731-9743. [3] Tabassum, f., islam, m. I., khan, r. T., & amin, m. R. (2020). Human face recognition with combination of dwt and machine Learning. Journal of king saud university-computer and information sciences. [4] Mellouk, w., & handouzi, w. (2020). Facial emotion recognition using deep learning: review and insights. Procedia computer science, 175, 689-694. [5] Hassouneh, a., mutawa, a. M., & murugappan, m. (2020). Development of a real-time emotion recognition system using facial expressions and eeg based on machine learning and deep neural network methods. Informatics in medicine unlocked, 20, 100372.

Copyright

Copyright © 2024 Dhyan Patel, Prof. Sonam Singh. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.