relationship between fatigue and safety risks in driving

relationship between fatigue and safety risks in driving

Introduction

Driving activities require a lot of focus and a great deal of mental strength. In addition, driving requires proper coordination of the movement of the hands and feet, moving gears and looking at other vehicles. Lack of attention can be deadly for drivers, pedestrians, and economic activity. Therefore, a country needs to have solid rules and regulations to maintain safety in the transportation of land. Also, the complexity and road characteristics can affect the driver’s attention. For example, in Indonesia, the existing road infrastructure does not mean much in one respect, and the number of vehicles produced in relation to road infrastructure [1] is not proportional. This is especially true in all major cities of developed and progressive countries. The traffic problem always keeps increasing from time to time. Even in small countries like Indonesia, estimated traffic fatality could be 40,000 per year in 2020 [2].

According to [3], there is a strong relationship between fatigue and safety risks in driving. Fatigue can be affected by health and sleep problems. Extreme fatigue can lead to the drowsiness of drivers who are considered to be the culprits of road accidents and can result in serious injuries, and a higher risk of death. In this regard, drowsiness is known as a lack of alertness or a lack of alertness that causes the driver to fall asleep while driving.

Facial Detection from Camera

The major work of machine vision starts from by acquiring the image data from the camera. After acquiring the image data, the next step is to find the facial feature and estimate the face area from the image acquired. Over the past decade, face detection and recognition have been one of the better and successful applications of popular fields of research and imagery, from intimate research to computer vision.

Analysis and understanding based on algorithms

Due to the intrinsic nature of the problem, computer vision is not only an area of ​​computer science research but also of the purpose of neuroscientific and psychological studies, mainly due to the general opinion that advances in computer image processing and understanding research Will provide. How our brain works and vice versa.

A general description of a facial recognition problem (in computer vision) can be compiled as follows: For a scene video or images, identify one or more people in the scene using a database of faces. [4]

Facial detection where an image is searched to find a face, then the image is processed, and the face is easily removed for identification.

Since 2002, face recognition with Intel’s Open Source Framework called Open CV can be quite easily and reliably done. This framework has an in-built face detector that works around 90-95% of clear photos of someone waiting for the camera. However, when looking at a person from an angle, it is usually difficult to detect his face, sometimes requiring 3D head poses. Also, the lack of proper image brightness can greatly increase the difficulty of face detection, or vice versa, or the image may be blurry, or the person may be wearing glasses, Etc.

Head Angle Estimation for Image Alignment

Facial modelling and analysis have long been an active research topic in computer vision [5] Larger datasets of faces [6] and various facial analysis problems have been effectively suggested over the years, such as facial recognition or identification of facial age [7] detection, and head angle estimation [8]. This article addresses the problem of head angle estimation, with many applications such as driver behaviour monitoring and human attention modelling. It can also be used to improve or provide additional information for other issues such as identity or identity recognition, expression recognition, or attention detection.

Detecting a head angle with a single image is a difficult problem. The angle of the head is a 3D vector that contains angles, pitch and roll angles. In order to estimate head angle from an image, it is necessary to learn the mapping between 2D and 3D space. Some methods utilize more modalities, such as depth information from 3D information [9] or video information over time. Depth images provide 3D information that is missing in 2D images. Videos capture the constant movement of human heads and provide additional information to help with angle estimation. However, learning to transient information is usually achieved through the high cost of a repetitive structure, while in-depth information often requires specific cameras that are not always available. Facial sign detection [10] is used to estimate the position of the head in most single frame angle estimation methods. However, it will count more and leads to bigger models. Therefore, not all these models are suitable for adoption on platforms with proper memory and computing resources.

Tiredness and Sleepiness During Driving

Naturally, fatigue can diminish attention, diminish information received, and decrease the ability to provide a response or assess a condition. This situation can lead to the wrong decision when driving. The worst can happen when the driver becomes drowsy and thus loses control of the sleeping vehicle. To prevent this scenario from occurring, we need to monitor driving activity to determine if the driver is asleep. This type of system can be used to create some warning alarms to alert the driver or send a signal to family relatives or friends. In fact, it is a challenge to detect a driver’s drowsiness due to the different ambient and lighting conditions (sleep apnea, talking, waking up, wearing glasses).

Problem Statement

Currently, the road safety is only conducting by the human intelligence and road sings. Its human mind that can take decision on road signs and the situation of the entire environment during driving. If the human mind is in the state of tiredness or sleepiness the then the driver cannot take the accurate decision during driving and result into the accidents and severe causalities of human lives.

Following are the problem statement currently facing during driving of a vehicle.

• No simple solution is available to find the human activeness level during driving
• The main cause of the accident during driving is to rotate the head angle towards other direction, and no proper solution is available in vehicles to find the accurate position of the head during driving.
• After finding the right angle of the face towards the font mirror of a vehicle the tiredness and sleepiness is a major cause of accidents and no real-time accurate solution is available to find the human tiredness and sleepiness level.
• Currently, there is no any caution system is available for the driver during driving for raising alertness level.

Research Objective

The main objective of this research is to define and develop such framework for the driving solution that can enhance the safe driving ability of the driver and enforce a driver to maintain his alertness and activeness during the driving moments.

Following is the research objectives that will be achieved during this study.

• To define and develop a simple solution with Machine Vision and Artificial Intelligence is available to find the human activeness level during driving
• To solve the main cause of the accident during driving by finding the right angle between driver and front mirror of the vehicle and if driver rotates his head towards other direction for more than a little moment, then create a caution system for the driver to reduce the accidents incidents.
• To find the tiredness and sleepiness level of a diver in real-time with an accurate solution during driving moments to reduce the accidental incidents.
• To define a caution system for the driver during driving for raising his alertness level.

Scope of the study

The scope of this study is related to the safety of humanity. Currently, the traffic accidents are increasing day by day. The human is keener to drive the vehicle by their own style. Still, the safety measurement is not directly handled by any vehicle apparatus or software that ensure the driver alertness and activeness during the driving. The main objective of this research is to define and develop such framework for the driving solution that can enhance the safe driving ability of the driver and enforce a driver to maintain his/her alertness and activeness during the driving moments.

This study’s implementation can be done in a various section like the public transport systems where a huge number of passengers are making their journey by a single bus or train. The life of all passengers is in the hand of the driver, and by utilizing the facility and advantage of our proposed model, the transport companies can ensure the safety of their passengers’ life by making the driver more alert and active during the driving situation.

The insurance companies can also use the benefits of our reach as the insurance companies have to pay insurance amount of claim for the accidents by the insured person. If it proves by our defined model that this accident is done by the carelessness of driver and he was in the situation of sleepiness and tiredness, then the insurance company doesn’t have to pay for their claim. By implementing this law, the insured person will drive his car more safely.

Research Methodology

The research mythology and research steps are following.

• First, this study will examine the thorough examine of currently available systems of machine vision to use a camera and detect the human face from images more accurately.
• Second, this study will find the solution of accurately alignment of a human face with defining the head angle towards the camera and front mirror of the vehicle.
• Third, after successful alignment of human face for best image fitting for detection, the next work is to find the human behaviour during driving.
• Fourth, by using the machine vision system, this study will find the facial features from ROI and do an analysis for finding the current situation of eyes that either the eyes are opened or closed.
• Fifth, by using the artificial intelligence, this study will define the activeness and sleepiness level of a driver during driving.
• Sixth, this study will define the alertness system for a driver that if he pays less attention to driving and rotates his head towards other direction or in a sleepiness state, then the system will create a high alert for him to pay full attenuation and remain active during driving moments.