Intelligent Driver Assistance Systems Research Articles

Driver Drowsiness Detection System Based on Eye State Analysis

Driver Drowsiness Detection System Based on Eye State Analysis

Synthesis of ISO 26262 and ISО 21448 methods to increase reliability of hazard analysis and risk assessment of intelligent driver assistance systems operation

Introduction (problem statement and relevance). Application of sensors to increase automation of vehicles is a growing global trend. However, as the number of these devices in vehicles increases, so does the number of risk factors caused by their application.The purpose of the study is to reduce the areas of unknown potentially dangerous situations of possible impact of these risks on a highly automated vehicle using a synthesis of the risk analysis methods.Methodology and research methods. This paper considers the system analysis methodologies set forth in GOST R ISO 26262, ISO 21448:2022.The scientific novelty is the interference of approaches to the analysis that are set forth in the standards.Practical significance. The results of application of the synthesized approach to risk analysis will form the basis of studies of the prospective intelligent driver assistance systems thus allowing the development of test programs for these systems within highly automated vehicles.

Enhancing Road Safety with Convolutional Neural Network Traffic Sign Classification

Recent computer vision and deep learning breakthroughs have improved road safety by automatically classifying traffic signs. This research uses CNNs to classify traffic signs to improve road safety. Autonomous vehicles and intelligent driver assistance systems require accurate traffic sign detection and classification. Using deep learning, we created a CNN model that can recognize and classify road traffic signs. This research uses a massive dataset of labeled traffic sign photos for training and validation. These CNN algorithms evaluate images and produce real-time predictions to assist drivers and driverless cars in understanding traffic signs. Advanced driver assistance systems, navigation systems, and driverless vehicles can use this technology to give drivers more precise information, improving their decision-making and road safety. Researcher optimized CNN model design, training, and evaluation metrics during development. The model was rigorously tested and validated for robustness and classification accuracy. The research also solves real-world driving obstacles like illumination, weather, and traffic signal obstructions. This research shows deep learning-based traffic sign classification can dramatically improve road safety. This technology can prevent accidents and enhance traffic management by accurately recognizing and interpreting traffic signs. It is also a potential step toward a safer, more efficient transportation system with several automotive and intelligent transportation applications. Road safety is a global issue, and CNN-based traffic sign classification can reduce accidents and improve driving. On filter 3, Convolutional Neural Network training accuracy reached 98.9%, while validation accuracy reached 88.23%.

High-Dynamic-Range Tone Mapping in Intelligent Automotive Systems

Intelligent driver assistance systems are becoming increasingly popular in modern passenger vehicles. A crucial component of intelligent vehicles is the ability to detect vulnerable road users (VRUs) for an early and safe response. However, standard imaging sensors perform poorly in conditions of strong illumination contrast, such as approaching a tunnel or at night, due to their dynamic range limitations. In this paper, we focus on the use of high-dynamic-range (HDR) imaging sensors in vehicle perception systems and the subsequent need for tone mapping of the acquired data into a standard 8-bit representation. To our knowledge, no previous studies have evaluated the impact of tone mapping on object detection performance. We investigate the potential for optimizing HDR tone mapping to achieve a natural image appearance while facilitating object detection of state-of-the-art detectors designed for standard dynamic range (SDR) images. Our proposed approach relies on a lightweight convolutional neural network (CNN) that tone maps HDR video frames into a standard 8-bit representation. We introduce a novel training approach called detection-informed tone mapping (DI-TM) and evaluate its performance with respect to its effectiveness and robustness in various scene conditions, as well as its performance relative to an existing state-of-the-art tone mapping method. The results show that the proposed DI-TM method achieves the best results in terms of detection performance metrics in challenging dynamic range conditions, while both methods perform well in typical, non-challenging conditions. In challenging conditions, our method improves the detection F2 score by 13%. Compared to SDR images, the increase in F2 score is 49%.

Machine learning based driver monitoring system: A case study for the Kayoola EVS

With the ever-growing traffic density, the number of road accidents has continued to increase. Finding solutions to reduce road accidents and improve traffic safety has become a top priority for Kiira Motors Corporation, a Ugandan state-owned automotive company. The company seeks to develop intelligent driver assistance systems for its market entry product, the Kayoola EVS bus. A machine learning-based driver monitoring system that would monitor driver drowsiness and send out an alarm in case drowsiness is detected has been developed in an attempt to reduce drowsiness-related accidents. The system consists of a camera positioned in such a way as to keep track of the driver's face. The camera is interfaced with a Raspberry Pi minicomputer which carries out the computations and analysis and when drowsiness is detected, an alarm is triggered. Dangerous driver behavior including distraction and fatigue has long been recognized as the main contributing factor in traffic accidents. This paper therefore presents the development of a driver monitoring system for the Kayoola Electric City Bus to address the increasing occurrences of road accidents. The machine learning-based driver monitoring system is designed to be non-intrusive with continuous real-time operation.

Research on Recognition of Road Hypnosis in the Typical Monotonous Scene.

Road traffic safety can be influenced by road hypnosis. Accurate detection of the driver's road hypnosis is a very important function urgently required in the driver assistance system. Road hypnosis recurs frequently in a certain period, and it tends to occur in a typical monotonous scene such as a tunnel or a highway. Taking the scene of a tunnel or a highway as a typical example, road hypnosis was studied through simulated driving experiments and vehicle driving experiments. A road hypnosis recognition model based on principal component analysis (PCA) and a long short-term memory network (LSTM) was proposed, where PCA was used to extract various parameters collected by the eye tracker, and the LSTM model was constructed to identify road hypnosis. The accuracy rates of 93.27% and 97.01% in simulated driving experiments and vehicle driving experiments were obtained. The proposed method was compared with k-nearest neighbor (KNN) and random forest (RF). The results showed that the proposed PCA-LSTM model had better performance. This paper provides a novel and convenient method to realize the driver's road hypnosis detection function of the intelligent driver assistance system in practical applications.

Intelligent Driver Assistance and Energy Management Systems of Hybrid Electric Autonomous Vehicles

Automotive companies continue to develop integrated safety, sustainability, and reliability features that can help mitigate some of the most common driving risks associated with autonomous vehicles (AVs). Hybrid electric vehicles (HEVs) offer practical solutions to use control strategies to cut down fuel usage and emissions. AVs and HEVs are combined to take the advantages of each kind to solve the problem of wasting energy. This paper presents an intelligent driver assistance system, including adaptive cruise control (ACC) and an energy management system (EMS), for HEVs. Our proposed ACC determines the desired acceleration and safe distance with the lead car through a switched model predictive control (MPC) and a neuro-fuzzy (NF) system. The performance criteria of the switched MPC toggles between speed and distance control appropriately and its stability is mathematically proven. The EMS intelligently control the energy consumption based on ACC commands. The results show that the driving risk is extremely reduced by using ACC-MPC and ACC-NF, and the vehicle energy consumption by driver assistance system based on ACC-NF is improved by 2.6%.

Development of Intelligent Driver Assistant System for Cognitive Cars

Automation in vehicles is essential in present day’s road conditions. Many of the accidents are happening due to wrong behaviour, drowsiness and human inattention. The vehicle drivers are expected to be more attentive during their driving period. As a normal human being, it is usual that sometimes there may be a lapse in attention due to many environmental factors. Therefore, automation in cars is really required to overcome many uncertainties faced in real-time scenarios. This paper explores various automation techniques used in vehicles. Cognitive cars proposed in the paper, functions based on machine learning Algorithms. These cars travel from source to destination without any human intervention. The main objectives included in the paper are lane detection, obstacle detection, road signs detection and traffic lights detection. In near future cognitive cars will be preferred for the modern communities due to its intelligent features and increased safety aspects. It will be an asset of the future generation.

A Real-Time Recognition System of Driving Propensity Based on AutoNavi Navigation Data.

Driving propensity is the driver’s attitude towards the actual traffic situation and the corresponding decision-making or behavior during the driving process. It is of great significance to improve the accuracy of safety early warning and reduce traffic accidents. In this paper, a real-time identification system of driving propensity based on AutoNavi navigation data is proposed. The main work includes: (1) A dynamic data acquisition method of AutoNavi navigation is proposed to obtain the time, speed and acceleration of the driver during the navigation process. (2) The dynamic data collection method of AutoNavi navigation is analyzed and verified through the dynamic data obtained in the real vehicle experiment. The principal component analysis method is used to process the experimental data to extract the driving propensity characteristics variables. (3) The fruit fly optimization algorithm combined with GRNN (generalized neural network) and the feature variable set are used to build a FOA-GRNN-based model. The results show that the overall accuracy of the model can reach 94.17%. (4) A driving propensity identification system is constructed. The system has been verified through real vehicle test experiments. This paper provides a novel and convenient method for building personalized intelligent driver assistance systems in practical applications.

Adaptive Cruise Predictive Control Based on Variable Compass Operator Pigeon-Inspired Optimization

A vehicle adaptive cruise system can control the speed and the safe distance between vehicles rapidly and effectively, which is an integral part of an intelligent driver assistance system. Adaptive cruise predictive control algorithms based on variable compass operator pigeon-inspired optimization (PIO) and PSO are proposed to improve the time response characteristics of multi-objective adaptive cruise system predictive control. Firstly, a longitudinal kinematic model of an adaptive cruise system was established and linearly discretized. Secondly, the multi-objective optimal cost function and parameter constraints were designed by integrating factors such as distance error, relative speed, acceleration and impact, and a mathematical model of the adaptive cruise predictive control optimization problem was constructed. Finally, PIO and PSO were used to solve the optimal control law for MPC and simulated by Matlab. The results show that the adaptive cruise system can reach a steady state quickly with the control laws of PIO or PSO. However, due to the global optimization and fast convergence characteristic, variable compass operator PIO has better time response characteristics.

An Intelligent Driving Assistance System Based on Lightweight Deep Learning Models

An intelligent driver assistance system is developed in this study, which is able to remind the drivers to turn on the head lights or wipers through situation recognition method when driving at night or on rainy days. Furthermore, it is also able to alarm the drivers based on the lightweight deep learning model and the distance estimation method when surrounding vehicles are too close. Experimental results show that the proposed methods and the chosen lightweight model in our proposed system obtain reliable performance and sufficient computational efficiency under limited computing resource. In conclude, our proposed system obtains high probability to be adopted for the development of advanced driver assistance systems (ADAS).

The choice of rational complex of active safety systems for smart cars

The problem of ensuring road safety affects all elements of the Driver-Car-Road-Environment system. Smart cars equipped with enough traffic assistants can significantly improve road safety. Active vehicle safety systems, including intelligent driver assistance systems and assistants, perform similar road safety functions. With all the variety of possibilities for equipping cars with systems complexes, the need arises to assess the feasibility and profitability of installing a particular complex of systems. For this, it is proposed to apply the methods of multi-criteria assessment. As a result of calculations, the best options for the sets of systems that most widely cover the road situation have been identified.

Description and tests of a control system for automatic emergency braking of a car with an algorithm for predicting the adhesion of automobile tires to the road surface

Every year the number of cars in the world is steadily growing, which in turn leads to an increase in road accidents. Russia, as the largest country by area in the world, has a number of traffic-related features that are not typical for other regions, including changing road and climatic conditions. The sharp increase in the number of accidents in the first hours after precipitation is especially noticeable, that indicates the difficulty to adapt to the changing traffic situation for vehicle drivers. More than 60% of collisions, according to the traffic police databases for 2018-2019, occur due to incorrect predictions by drivers of the braking distance of their cars. As a rule, this can happen due to the inexperience of the driver, incorrect assessment of the current adhesion properties of the roadway, distraction to a mobile phone, the climate or multimedia system of the car, movement with insufficient visibility. Preventing collisions occurring in such scenarios is the main task of an intelligent driver assistance system such as the automatic emergency braking system. The main part of this system is an adequate assessment of dynamically changing road and climatic conditions and warning the driver of a possible collision hazard, as well as automatic action on the vehicle’s brake drive to prevent a collision. The article describes control algorithms for assessing and predicting dangerous situations.

Braking intensity recognition with optimal K-means clustering algorithm

Recognizing a driver’s braking intensity plays a pivotal role in the development of modern driver assistance and energy management systems; and therefore, it is especially important to autonomous and electric vehicles. This paper aims at developing a system for recognizing a driver’s braking intensity based on the pressure produced in the brake master cylinder. In this regard, a model-based, synthetic data generation concept is used to generate the training dataset. This technique involves two closed-loop controlled models: an upper-level longitudinal vehicle dynamics model, and a lower-level brake hydraulic, dynamic model. The adaptive particularly tunable fuzzy particle swarm optimization algorithm is recruited to solve the optimal K-means clustering problem and to determine the best number of clusters and centroids into which the brake pressure data are to be divided. The obtained results reveal that the brake pressure data for a vehicle traveling the new European driving cycle can be best partitioned into two clusters. A driver’s braking intensity may, therefore, be clustered as moderate or intensive. With the ability to automatically recognize a driver’s pedal feel, the system developed in this research could be implemented in intelligent driver assistance systems as well as in electric vehicles equipped with intelligent, electromechanical brake boosters.

Russian national non-commercial vehicle safety rating system RuNCAP

The article is devoted to the need to introduce an independent rating system for assessing the safety of cars RuNCAP in the Russian Federation. The introduction of this system will make it possible to more fully take into account the Russian operating conditions when assigning ratings to passive and active safety. RuNCAP will contribute to the wider adoption of intelligent driver assistance systems and their efficiency. The development of RuNCAP is proposed to be carried out taking into account the analysis of the features of the accident statistics for the Russian Federation. The article also presents the development strategy of RuNCAP for the coming years.

Training a thin and shallow lane detection network with self-knowledge distillation

With modern science and technology development, vehicles are equipped with intelligent driver assistant systems, of which lane detection is a key function. These complex detection structures (either wide or deep) are investigated to boost the accuracy and overcome the challenges in complicated scenarios. However, the computation and memory storage cost will increase sharply, and the response time will also increase. For resource-constrained devices, lane detection networks with a low cost and short inference time should be implemented. To get more accurate lane detection results, the large (deep and wide) detection structure is framed for high-dimensional and highly robust features, and deep supervision loss is applied on different resolutions and stages. Despite the high-precision advantage, the large detection network cannot be used for embedded devices directly because of the demand for memory and computation. To make the network thinner and lighter, a general training strategy, called self-knowledge distillation (SKD), is proposed. It is different from classical knowledge distillation; there are no independent teacher-student networks, and the knowledge is distilled itself. To evaluate more comprehensively and precisely, a new lane data set is collected. The Caltech Lane date set and TuSimple lane data set are also used for evaluation. Experiments further prove that a small student network and large teacher network have a similar detection accuracy via SKD, and the student network has a shorter inference time and lower memory usage. Thus it can be applied for resource-limited devices flexibly.

Detection of driving actions on steering wheel using triboelectric nanogenerator via machine learning

Abstract The sensor-based early recognition of driver’s driving actions on steering wheels is a complementary means of Intelligent Driver Assistance Systems (IDAS), which can help prevent traffic accidents. In this paper, using triboelectric nanogenerator (TENG) as sensor for detecting driver’s steering actions is studied. Two driving simulator based experiments are designed and conducted. First, response speed of three sensors (driving simulator, camera and TENG) is quantified and compared. Then, a machine learning algorithm is designed and trained to detect three types of steering actions. Using this algorithm, electrical signals from TENGs can be used to detect driver’s steering actions. Our results show that TENG has the fastest response speed statistically. The trained algorithm has an accuracy of 92.0% in test dataset. This study may demonstrate the potential in using TENG as sensor for driver’s steering action detection.

Advances in Experimental Vehicle Soiling Tests

<div class="section abstract"><div class="htmlview paragraph">The field of vision of the driver during wet road conditions is essential for safety at all times. Additionally, the safe use of the increasing number of sensors integrated in modern cars for autonomous driving and intelligent driver assistant systems has to be ensured even under challenging weather conditions. To fulfil these requirements during the development process of new cars, experimental and numerical investigations of vehicle soiling are performed. This paper presents the surface contamination of self- and foreign-soiling tested in the wind tunnel. For these type of tests, the fluorescence method is state-of-the-art and widely used for visualizing critical areas. In the last years, the importance of parameters like the contact angle have been identified when designing the experimental setup. In addition, new visualization techniques have been introduced. Based on these developments the complete process using the fluorescence method was evaluated and updated to the available technology. Various factors in the experimental setup were identified to improve the reliability and a resulting high repeatability of the experiments between different test sessions is shown. Additionally, new visualization techniques are introduced for enhanced data evaluation which allows an easier identification of possible measures. Due to the increased repeatability, it is now possible to calculate the absolute difference between two analyzed vehicle configurations. For the side mirror glass, the quality of the results is improved by image detection methods. In the end it is shown that these modifications to the process are helpful in evaluating the results of computational fluid dynamics.</div></div>

The employment of multi-criteria assessment methods in the study of road safety

Intelligent driver assistance systems and traffic assistants are becoming more technological and widespread. Due to the great diversity, there is a need to compare their effectiveness in different conditions and choose a set of systems that is the most efficient and inexpensive. The article considers possible comparison criteria. Some features that are worth considering when conducting a study are identified. The analysis of statistics on road traffic accidents.

Intelligent driver assistance systems as factor of transportation safety assurance

The article discusses measures aimed at preventing road traffic accidents (RTAs). The main causes and conditions contributing to RTAs are investigated, specific measures are proposed to reduce the accident rate and its socially dangerous consequences for the country. The analysis of RTA causes shows that most of them are due to the driver’s fault. The concept of solving the issue of safety of road users is substantiated not in terms of an already occurred collision but in terms of its prevention. In order to prevent possible accidents on roads, as well as to ensure the safety of drivers and passengers we suggest equipping vehicles with the ADAS Eagle Vision intelligent driver assistance system. The system monitors the distance between vehicles, timely warning of pedestrians on the roadway. A warning of violations and danger comes in the form of sound signals and displayed visually on the LCD display. Installation of this system will reduce the number of RTAs occurring due to the driver’s fault. Based on the calculations performed, we can conclude that it is expedient to use the ADAS Eagle Vision intelligent driver assistance system.