Search

Distracted driving is a significant threat to road safety, causing numerous accidents every year. Driver distraction detection systems offer a promising solution by alerting the driver to refocus on the primary driving task. Even with increasing vehicle automation, human drivers must remain alert, especially in partially automated vehicles where they may need to take control in critical situations. In this work, an AI-based distraction detection model is developed that focuses on improving classification performance using a long short-term memory (LSTM) network. Unlike traditional approaches that evaluate individual frames independently, the LSTM network captures temporal dependencies across multiple time steps. In addition, this study investigated the integration of vehicle sensor data and an inertial measurement unit (IMU) to further improve detection accuracy. The results show that the recurrent LSTM network significantly improved the average F1 score from 71.3% to 87.0% compared to a traditional vision-based approach using a single image convolutional neural network (CNN). Incorporating sensor data further increased the score to 90.1%. These results highlight the benefits of integrating temporal dependencies and multimodal inputs and demonstrate the potential for more effective driver distraction detection systems that can improve road safety. [ABSTRACT FROM AUTHOR]

Among the sensors necessary to equip vehicles with an autonomous driving system, there is a tacit agreement that cameras and some type of radar would be essential. The ability of radar to spatially locate objects (pedestrians, other vehicles, trees, street furniture, and traffic signs) makes it the most economical complement to the cameras in the visible spectrum in order to give the correct depth to scenes. From the echoes obtained by the radar, some data fusion algorithms will try to locate each object in its correct place within the space surrounding the vehicle. In any case, the usefulness of the radar will be determined by several performance parameters, such as its average error in distance, the maximum errors, and the number of echoes per second it can provide. In this work, we have tested experimentally the AWR1843 MIMO radar from Texas Instruments to measure those parameters. [ABSTRACT FROM AUTHOR]

The transportation of goods by road is crucial in Tanzania's Central and North-South Corridors. However, challenges such as truck congestion and road accidents are affecting the efficiency of these routes. Road crashes are prevalent in low- and middle-income countries, with Africa experiencing an exceptionally high rate. This study examines the opinions of Tanzanian truck drivers on the effectiveness of Advanced Driver Assistance Systems in reducing road safety risks. A discriminant analysis was conducted to assess the awareness and use of these systems among different driver experienced groups. The results highlight the importance of improving infrastructure, ensuring vehicle safety standards, providing comprehensive driver training, and integrating innovative Intelligent Transport Systems to address road safety issues. In conclusion, the study provides valuable insights for policymakers and stakeholders to strengthen road safety measures in Tanzania, facilitating smoother road freight transport operations and promoting economic growth. [ABSTRACT FROM AUTHOR]

Close following to lead vehicles is associated with increased risk of rear-end crashes in road traffic. One way to reduce instances of close following is through increased use of the Advanced Driver Assistance System (ADAS) Adaptive Cruise Control (ACC), which is designed to adjust vehicle speed to maintain a safe time headway. Since the activation of ACC is driver-initiated, there is a need to influence the propensity of drivers to use the function. This research aimed to explore whether in-vehicle nudging interventions could be effective for this purpose. A field trial was conducted to consecutively assess the effects of two nudges on drivers' utilization of ACC, compared to baseline usage. Exposing the participants (n = 49) to the first ambient design nudge resulted in a 46% increase in ACC usage on average. Following the introduction of the second nudge (a competitive leaderboard nudge), the average increase among participants (n = 48) during the complete treatment period reached 61%. The changes in ACC utilization varied between individual drivers, highlighting the need to monitor behavioral outcomes of nudges and adapt them when needed. In conclusion, this research shows that utilizing in-vehicle nudging is a promising approach to increase the use of vehicle functions contributing to improved traffic safety. [ABSTRACT FROM AUTHOR]

Acquiring high-quality images in diverse lighting conditions is challenging, particularly in the constantly evolving driving landscape. From the harsh brilliance of daylight to the subtle shadows of nighttime scenarios, the complexity of capturing accurate images intensifies. Therefore, addressing exposure issues across various lighting conditions is imperative. This study introduces a novel multi-task end-to-end dynamic illumination adjustment approach termed CPGA-DIA. By incorporating feature selection and gamma factor correction techniques, our method effectively handles extreme variations in light conditions for both low-light image enhancement and exposure correction. Comprehensive experimental evaluations demonstrate superior performance to existing processes, showcasing our method's adaptability and robustness in real-world driving scenarios. Furthermore, our approach extends its capabilities to include the control and estimation of explainable factors related to changing lightness, enabling our model to understand better the environmental conditions that influence image perception and clarity. This research contributes to advancing image capture methodologies, particularly in dynamic lighting conditions, with promising implications for applications ranging from photography to various other fields. [ABSTRACT FROM AUTHOR]

Abstract Camera‐based Driver Monitoring Systems (DMS) have the potential to exploit eye tracking correlates of alcohol intoxication to detect drunk driving. This study investigates how glance, blink, saccade, and fixation metrics are affected by alcohol, and whether possible effects remain stable across three different camera setups, as well as when the driver is out‐of‐the‐loop during level 4 automated driving (Wizard‐of‐Oz setup). Thirty‐five participants drove on a test track first sober and then with increasing intoxication levels reaching a breath alcohol concentration (BrAC) of 1‰. Linear Mixed‐Effects Regression analyses showed that with increasing intoxication levels, eye blinks became longer and slower, glances and fixations became fewer and longer, and more attention was directed to the road area, at the expense of more peripheral areas. Fixation and blink metrics were more robust to changes in automation mode, whereas glance‐based metrics were highly context dependent. Not all effects of alcohol intoxication could be measured with all eye tracking setups, where one‐camera systems showed lower data availability and higher noise levels compared to a five‐camera system. This means that lab findings based on higher quality eye tracking data might not be directly applied to production settings because of hardware limitations.

Among the sensors necessary to equip vehicles with an autonomous driving system, there is a tacit agreement that cameras and some type of radar would be essential. The ability of radar to spatially locate objects (pedestrians, other vehicles, trees, street furniture, and traffic signs) makes it the most economical complement to the cameras in the visible spectrum in order to give the correct depth to scenes. From the echoes obtained by the radar, some data fusion algorithms will try to locate each object in its correct place within the space surrounding the vehicle. In any case, the usefulness of the radar will be determined by several performance parameters, such as its average error in distance, the maximum errors, and the number of echoes per second it can provide. In this work, we have tested experimentally the AWR1843 MIMO radar from Texas Instruments to measure those parameters.

AbstractCommunicating effectively with autonomous vehicles requires contextualized visual and auditory cues to ensure clear message delivery. Evaluating the user experience involves assessing which types of information can be safely reacted to without additional monitoring and how it is presented. Validating the visual and auditory cues supports the driver’s course of action. This study investigates message types and preferred modalities of driver-to-driver communication via vehicle-to-everything (V2X) in advanced driver assistance systems technologies for autonomous and manual driving and proposes efficient ways to respond to event situations. Four modalities, including baseline and three message types with different information were proposed to investigate the information required by drivers. Results indicate that providing notifications during autonomous driving is more helpful and less workload-intensive than during manual driving. Most notifications were highly visible and easy to recognize. Although behavioral messages in both autonomous and manual driving enhance usability, providing advice and behavioral messages is safer for autonomous driving. Designing V2X notification information based on future events is vital because of its highly pragmatic nature. [ABSTRACT FROM AUTHOR]

Camera‐based Driver Monitoring Systems (DMS) have the potential to exploit eye tracking correlates of alcohol intoxication to detect drunk driving. This study investigates how glance, blink, saccade, and fixation metrics are affected by alcohol, and whether possible effects remain stable across three different camera setups, as well as when the driver is out‐of‐the‐loop during level 4 automated driving (Wizard‐of‐Oz setup). Thirty‐five participants drove on a test track first sober and then with increasing intoxication levels reaching a breath alcohol concentration (BrAC) of 1‰. Linear Mixed‐Effects Regression analyses showed that with increasing intoxication levels, eye blinks became longer and slower, glances and fixations became fewer and longer, and more attention was directed to the road area, at the expense of more peripheral areas. Fixation and blink metrics were more robust to changes in automation mode, whereas glance‐based metrics were highly context dependent. Not all effects of alcohol intoxication could be measured with all eye tracking setups, where one‐camera systems showed lower data availability and higher noise levels compared to a five‐camera system. This means that lab findings based on higher quality eye tracking data might not be directly applied to production settings because of hardware limitations. [ABSTRACT FROM AUTHOR]

When a ground vehicle runs at high speeds, even a slight excess in the wheel steering angle can immediately cause the vehicle to slide sideways and lose control. In this study, we propose an active safety control system designed to address emergency situations where the driver applies excessive steering input and the vehicle speed varies significantly during control. The system combines the direct yaw moment (DYM) method with a steering saturation scheme that prevents excessive driver steering input from adversely influencing the front-wheel steering. Consequently, the control system allows the DYM to focus more on other stabilization tasks and maintain tire/road friction within its workable linear range. The implementation relies on a reference steering angle and a reference vehicle state, derived from a linear vehicle model considering tire/road friction limitations. When the driver's steering angle and the system state deviate from these reference values, the control system intervenes by applying both the steering saturation scheme and DYM method. This ensures the front-wheel steering angle and system state remain close to the reference values. The control strategy is developed using the polytopic Linear Parameter Varying (LPV) technique and Linear Matrix Inequality (LMI) to account for the changes in vehicle speed. It is further enhanced with an input saturation technique based on a high-gain approach, which improves control utilization and system response during emergency situations. The advantages of the proposed control strategy are demonstrated through simulation results. [ABSTRACT FROM AUTHOR]

Prediction of a driver's braking intention enables the advanced driver assistance system (ADAS) to intervene in the braking system as early as possible, which may shorten braking distance and improve driving safety. This paper proposes a novel deep learning model called LSTM-CNN-Attention that combines a long short-term memory (LSTM) neural network, convolutional neural network (CNN), and Attention mechanism for extracting spatiotemporal features of multi-sensor data to improve prediction accuracy. The proposed model inherits both temporal and spatial feature extraction abilities from LSTM and CNN. The LSTM-CNN-Attention model has a parallel architecture, which enhances the feature extraction ability of the model for multi-sensor time series data and improves the prediction accuracy of the driver's braking intention before the braking action. Furthermore, a driving simulator is set up to sample driving data for training and evaluating the proposed method. According to the results of the experiment, the model obtains up to 3.16% higher accuracy than the baseline models such as LSTM, CNN, and bidirectional LTSM (Bi-LSTM). Additionally, the influence of sliding window size and prediction horizon on the performance of the method is investigated. A method of tuning hyperparameters using the genetic algorithm is presented. The results demonstrate that the prediction accuracy increases by about 2% after being optimized by GA. [ABSTRACT FROM AUTHOR]

The acquisition and analysis of high-quality sensor data constitute an essential requirement in shaping the development of fully autonomous driving systems. This process is indispensable for enhancing road safety and ensuring the effectiveness of the technological advancements in the automotive industry. This study introduces the Interaction of Autonomous and Manually Controlled Vehicles (IAMCV) dataset, a novel and extensive dataset focused on intervehicle interactions. The dataset, enriched with a sophisticated array of sensors such as lidar, cameras, inertial measurement unit/Global Positioning System, and vehicle bus data acquisition, provides a comprehensive representation of real-world driving scenarios that include roundabouts, intersections, country roads, and highways, recorded across diverse locations in Germany. Furthermore, the study shows the versatility of the IAMCV dataset through several proof-of-concept use cases. First, an unsupervised trajectory clustering algorithm illustrates the dataset’s capability in categorizing vehicle movements without the need for labeled training data. Second, we compare an online camera calibration method with the Robot Operating System-based standard, using images captured in the dataset. Finally, a preliminary test employing the YOLOv8 object-detection model is conducted, augmented by reflections on the transferability of object detection across various lidar resolutions. These use cases underscore the practical utility of the collected dataset, emphasizing its potential to advance research and innovation in the area of intelligent vehicles. [ABSTRACT FROM AUTHOR]

Affective human–vehicle interaction of intelligent cockpits is a key factor affecting the acceptance, trust, and experience for intelligent connected vehicles. Driver emotion detection is the premise of realizing affective human–machine interaction. To achieve accurate and robust driver emotion detection, we propose a novel brain-inspired framework for on-road driver emotion detection using facial expressions. Then, we conduct driver emotion data collection in an on-road context. We develop a data annotation tool, annotate the collected data, and obtain the RoadEmo dataset, a dataset of facial expressions and road scenarios under the driver’s emotional driving. Finally, we validate the detection accuracy of the proposed framework. The experiment results show that our proposed framework achieves excellent detection performance in the on-road driver emotion detection task and outperforms existing frameworks. [ABSTRACT FROM AUTHOR]

Good data quality is crucial for any data-driven system's effective and safe operation. For critical safety systems, the significance of data quality is even higher since incorrect or low-quality data may cause fatal faults. However, there are challenges in identifying and managing data quality. In particular, there is no accepted process to define and continuously test data quality concerning what is necessary for operating the system. This lack is problematic because even safety-critical systems become increasingly dependent on data. Here, we propose a Candidate Framework for Data Quality Assessment and Maintenance (CaFDaQAM) to systematically manage data quality and related requirements based on design science research. The framework is constructed based on an advanced driver assistance system (ADAS) case study. The study is based on empirical data from a literature review, focus groups, and design workshops. The proposed framework consists of four components: a Data Quality Workflow, a List of Data Quality Challenges, a List of Data Quality Attributes, and Solution Candidates. Together, the components act as tools for data quality assessment and maintenance. The candidate framework and its components were validated in a focus group. [ABSTRACT FROM AUTHOR]

Abstract An object detection framework using thermal infrared (TIR) cameras is proposed to meet the needs of an advanced driver assistance system (ADAS) operating at night‐time and in low‐visibility conditions. The proposed detection framework, referred to as TIR‐YOLO‐ADAS, is an improvement of YOLOX for TIR object detection in ADAS. First, to address the disadvantages of TIR objects, the part of the attention mechanism is designed to enhance the discriminative ability of feature maps in the spatial and channel dimensions. Second, a focal loss function is used as the confidence loss function to enable the framework to focus on detection tasks of difficult, misclassified targets in the process of network training. The results of the ablation experiment on the Forward‐looking infrared (FLIR) thermal ADAS dataset indicate that the proposed framework significantly improves the performance of TIR object detection. Comparative experimental results further show that TIR‐YOLO‐ADAS performs favourably when compared with three representative detection algorithms. To evaluate the practicality and feasibility of the proposed framework in various applications, a qualitative assessment in real road scenarios was conducted. The experimental results confirm that the proposed framework performs promisingly and could be integrated into vehicle platforms as an ADAS module.

Situation awareness (SA) describes an individual's understanding of their surroundings and actions in the near future based on the individual's comprehension and understanding of the surrounding inputs. SA measurements can be applied to improve system performance or human effectiveness in many fields of study, including driving. However, in some scenarios drivers might need to drive in unfamiliar traffic regulations (UFTRs), where the traffic rules and vehicle configurations are a bit different from what the drivers are used to under familiar traffic regulations. Such driving conditions require drivers to adapt their attention, knowledge, and reactions to safely reach the destination. This ability is influenced by the degree of handedness. In such tasks, mixed-/left-handed people show better performance than strong right-handed people. This paper aims to explore the influence of the degree of handedness on SA when driving under UFTRs. We analyzed the SA of two groups of drivers: strong right-handed drivers and mixed-/left-handed drivers. Both groups were not familiar with driving in keep-left traffic regulations. Using a driving simulator, all participants drove in a simulated keep-left traffic system. The participants' SA was measured using a subjective assessment, named the Participant Situation Awareness Questionnaire PSAQ, and performance-based assessment. The results of the study indicate that mixed-/left-handed participants had significantly higher SA than strong right-handed participants when measured by performance-based assessment. Also, in the subjective assessment, mixed-/left-handed participants had significantly higher PSAQ performance scores than strong right-handed participants. The findings of this study suggest that advanced driver assistance systems (ADAS), which show improvement in road safety, should adapt the system functionality based on the driver's degree of handedness when driving under UFTRs. [ABSTRACT FROM AUTHOR]

Summary: This paper presents an extended version of Deeper, a search‐based simulation‐integrated test solution that generates failure‐revealing test scenarios for testing a deep neural network‐based lane‐keeping system. In the newly proposed version, we utilize a new set of bio‐inspired search algorithms, genetic algorithm (GA), (μ+λ) and (μ,λ) evolution strategies (ES), and particle swarm optimization (PSO), that leverage a quality population seed and domain‐specific crossover and mutation operations tailored for the presentation model used for modeling the test scenarios. In order to demonstrate the capabilities of the new test generators within Deeper, we carry out an empirical evaluation and comparison with regard to the results of five participating tools in the cyber‐physical systems testing competition at SBST 2021. Our evaluation shows the newly proposed test generators in Deeper not only represent a considerable improvement on the previous version but also prove to be effective and efficient in provoking a considerable number of diverse failure‐revealing test scenarios for testing an ML‐driven lane‐keeping system. They can trigger several failures while promoting test scenario diversity, under a limited test time budget, high target failure severity, and strict speed limit constraints. [ABSTRACT FROM AUTHOR]

An object detection framework using thermal infrared (TIR) cameras is proposed to meet the needs of an advanced driver assistance system (ADAS) operating at night‐time and in low‐visibility conditions. The proposed detection framework, referred to as TIR‐YOLO‐ADAS, is an improvement of YOLOX for TIR object detection in ADAS. First, to address the disadvantages of TIR objects, the part of the attention mechanism is designed to enhance the discriminative ability of feature maps in the spatial and channel dimensions. Second, a focal loss function is used as the confidence loss function to enable the framework to focus on detection tasks of difficult, misclassified targets in the process of network training. The results of the ablation experiment on the Forward‐looking infrared (FLIR) thermal ADAS dataset indicate that the proposed framework significantly improves the performance of TIR object detection. Comparative experimental results further show that TIR‐YOLO‐ADAS performs favourably when compared with three representative detection algorithms. To evaluate the practicality and feasibility of the proposed framework in various applications, a qualitative assessment in real road scenarios was conducted. The experimental results confirm that the proposed framework performs promisingly and could be integrated into vehicle platforms as an ADAS module. [ABSTRACT FROM AUTHOR]

This research aims to develop robust, resilient, and user-friendly over-the-air (OTA) software update mechanisms for advanced driver assistance systems (ADAS) in modern vehicles. The study conducts a comprehensive review on existing OTA update approaches, identifies security threats and vulnerabilities, and proposes an algorithm that incorporates cryptographic techniques, secure authentication, firmware validation, and rollback protection. The methodology utilized in this study involves secure package generation, distribution, installation, and monitoring, as well as adhering to automotive cybersecurity standards, such as ISO/SAE 21434 and UNECE WP.29. Extensive evaluation of OTA systems is important as it helps to assess their resilience against adversarial attacks, ensures regulatory compliance, and optimizes usability through user studies. The results demonstrate successful execution of secure OTA update mechanisms, preventing risks, fostering reliability and safety of ADAS software updates. The case studies highlight real-world deployments, best practices, and the effectiveness of the developed solution in improving automotive cybersecurity and functional safety. [ABSTRACT FROM AUTHOR]

The increasing number of vehicles on the road places high demands for autonomous driving in future intelligent transportation systems (ITSs). To realize automated services in ITSs, high-accuracy positioning is increasingly important, while positioning simultaneity for high-mobility vehicles should be guaranteed. Existing positioning systems have many constraints in locating vehicles in urban environments, making it difficult to meet the performance requirements. With the development of 6G, massive numbers of high-frequency base stations (BSs) are expected to be deployed to ensure connectivity and benefit positioning performance. However, high-frequency signals have severe distance attenuations and strong sensitivity to the non-line-of-sight (NLoS) environment that is prevalent in cities. Intelligent reflecting surface (IRS), with its ability to control signal reflections and create beyond-LoS communications, is a promising technology to improve positioning performance in urban vehicular networks. In this article, we discuss the existing positioning systems and explain the potentials of IRS for future BS-enabled positioning systems. We also provide an example to illustrate how to optimize the performance of IRS-aided vehicle positioning systems. Furthermore, we highlight important issues and give future directions. [ABSTRACT FROM AUTHOR]

Advanced driver assistance systems (ADASs) have recently gained popularity as they assist vehicle operators in staying within safe boundaries, helping them thereby to prevent possible collisions. However, despite their success and development, most ADAS use common and deterministic warning thresholds for all drivers in all driving environments. This may occasionally lead to the issuance of annoying inadequate warnings, due to the possible differences between drivers, the changing environments and driver statuses, thus reducing their acceptance and effectiveness. To fill this gap, this paper proposes adaptive algorithms for commonly used warnings based on real-time traffic environments and driver status including distraction and fatigue. We proposed adaptive algorithms for headway monitoring, illegal overtaking, over-speeding, and fatigue. The algorithms were then tested using a driving simulator. Results showed that the proposed adaptive headway warning algorithm was able to automatically update the headway warning thresholds and that, overall, the proposed algorithms provided the expected warnings. In particular, three or four different warning types were designed to distinguish different risk levels. The designed real-time intervention algorithms can be implemented in ADAS to provide warnings and interventions tailored to the driver status to further ensure driving safety. [ABSTRACT FROM AUTHOR]

Sight obstructions along road curves can lead to a crash if the driver is not able to stop the vehicle in time. This is a particular issue along curves with limited available sight, where speed management is necessary to avoid unsafe situations (e.g., driving off the road or invading the other traffic lane). To solve this issue, we proposed a novel intelligent speed adaptation (ISA) system for visibility, called V-ISA (intelligent speed adaptation for visibility). It estimates the real-time safe speed limits based on the prevailing sight conditions. V-ISA comes with three variants with specific feedback modalities (1) visual and (2) auditory information, and (3) direct intervention to assume control over the vehicle speed.Here, we investigated the efficiency of each of the three V-ISA variants on driving speed choice and lateral behavioural response along road curves with limited and unsafe available sight distances, using a driving simulator. We also considered curve road geometry (curve direction: rightward vs. leftward). Sixty active drivers were recruited for the study. While half of them (experimental group) tested the three V-ISA variants (and a V-ISA off condition), the other half always drove with the V-ISA off (validation group). We used a linear mixed-effect model to evaluate the influence of V-ISA on driver behaviour.All V-ISA variants were efficient at reducing speeds at entrance points, with no discernible negative impact on driver lateral behaviour. On rightward curves, the V-ISA intervening variant appeared to be the most effective at adapting to sight limitations. Results of the current study implies that V-ISA might assist drivers to adjust their operating speed as per prevailing sight conditions and, consequently, establishes safer driving conditions.

Accurate distance estimation is crucial for ensuring the safety and efficiency of intelligent and future autonomous mobility. In this paper, we propose a method for real-time estimation of visible lane distance based on image data only from a single camera and an existing road lane marking estimation system. By leveraging camera intrinsic and height above the road, we utilize lane parallelism and geometric properties of frames from a camera to compute the distance to the furthest visible lane point. Intelligent recalculations of horizon level are performed for every frame, accommodating variations encountered during uphill or downhill driving scenarios. Moreover, the proposed solution is designed to be theoretically adaptable to gauge the distance to any detected object on the camera frame, not only road lanes. We evaluate the effectiveness of our method on a dataset created for this purpose, assessing estimation performance through comparison with ground truth data obtained from high-precision GPS measurements. The results indicate comparable estimation performance with other works. Coupled with the capability for real-time implementation in vehicles, we underscore the potential of our approach for advanced driver assistance systems (ADAS) in various aspects, such as safe driving speeds, obstacle avoidance, or collision prevention. This research contributes to the ongoing efforts in developing robust and practical solutions for autonomous mobility applications.

This survey aims to provide an overview of various methods for generating data-driven test scenarios for assessing automated driving systems (ADSs). The survey updates the overall process of scenario generation and categorizes the current methods using a systematic literature review of 64 studies identified between 2017 and 01/2023. Overall, we demonstrate that the data-driven scenario generation process should be updated by another process step, scenario fusion, leading to seven process steps: 1) scope definition, 2) primary data source selection, 3) primary data collection, 4) scenario identification, 5) scenario fusion, 6) scenario generation, and 7) scenario evaluation. “Scenario fusion” aims to fuse scenarios identified in different data sources for a better coverage of the ADSs’ operational design domains (ODDs) and a more comprehensive scenario description. Moreover, we introduce an improved definition for the representativity of test scenario catalogs, which helps improve the collection of traffic data using sampling plans. Also, we show that real driving and police accident data are the most commonly used data input sources. Besides, we illustrate that the ODD is often not defined. Finally, we discuss that the standardization of test scenario generation is difficult because most methods do not address specific ADSs and test environments, and do not provide standardized interfaces. Overall, we recommend comparing existing approaches using the same input data and researching the mutual supplementation of the existing methods. Finally, pre-defined case studies, further standardized terminology, and standards for test execution and evaluation can help speed up the standardization process.

Automated driving systems should be able to avoid road traffic accidents and drive more safely than human drivers do. Test scenarios derived from real-world data such as police accident data can help assess the safety performance of automated driving systems. In many countries, the police collect data and information about nearly every accident, resulting in a representative sample. However, the collected accident data often do not contain the exact conflicts that cause an accident. Therefore, we estimated the globally known three-digit accident type for German police accident data, describing accident-causing conflicts. The data supplemented by the estimated three-digit accident type can then be used in subsequent test scenario generation. Accordingly, this study presents the first classification model for predicting 30 types of turning accidents. We tested a CatBoost model and a large-language model called BERT, using different feature sets and model designs. Overall, the CatBoost model performed best when using accident descriptions and non-text features such as collision type. Anomaly detection performed before model training revealed additional knowledge-driven miscoding in the police data collection. In conclusion, the model can predict common accident types, such as left-turns with an oncoming straight-ahead driver. In contrast, the model fails to predict rare accident types, such as left-turns (with a lit arrow sign) with oncoming traffic. Future studies should focus on optimizing the developed model, handling data imbalances, and building models using police accident data from other countries.

Scenario-based testing will help to validate automated driving systems (ADS) and establish safer road traffic. To date, most data-driven test scenario generation methods rely primarily on one data source such as police accident data (PD), naturalistic driving studies, or video-based traffic observations (VOs). However, none of these data sources perfectly satisfies all the layers of the six-layer model for the description of test scenarios. Moreover, not all available data sources cover the same location and period of time. Therefore, we fused information from 1,648 scenarios extracted from a German VO with information from 74 scenarios extracted from German PD into a comprehensive new PD* database. Finally, PD* consisted of 74 accident scenarios extended, for example, by variables containing the dynamic information of the VO scenarios. Thus, PD* contained more than 350 variables, whereas PD contained only 269 and VO only 122 variables. For fusion, we followed the Find-Unify-Synthesize-Evaluation (FUSE) for Representativity (FUSE4Rep) process model using statistical matching. Subsequently, we derived three logical scenarios from PD* to test an autonomous emergency braking system (AEB) in a stochastic traffic simulation incorporating driver-behavior models. The quality of the fusion itself was satisfactory, and we propose improving the VO data collection process and observation time to obtain even better results.

Distracted driving is a significant threat to road safety, causing numerous accidents every year. Driver distraction detection systems offer a promising solution by alerting the driver to refocus on the primary driving task. Even with increasing vehicle automation, human drivers must remain alert, especially in partially automated vehicles where they may need to take control in critical situations. In this work, an AI-based distraction detection model is developed that focuses on improving classification performance using a long short-term memory (LSTM) network. Unlike traditional approaches that evaluate individual frames independently, the LSTM network captures temporal dependencies across multiple time steps. In addition, this study investigated the integration of vehicle sensor data and an inertial measurement unit (IMU) to further improve detection accuracy. The results show that the recurrent LSTM network significantly improved the average F1 score from 71.3% to 87.0% compared to a traditional vision-based approach using a single image convolutional neural network (CNN). Incorporating sensor data further increased the score to 90.1%. These results highlight the benefits of integrating temporal dependencies and multimodal inputs and demonstrate the potential for more effective driver distraction detection systems that can improve road safety.

Sight obstructions along road curves can lead to a crash if the driver is not able to stop the vehicle in time. This is a particular issue along curves with limited available sight, where speed management is necessary to avoid unsafe situations (e.g., driving off the road or invading the other traffic lane). To solve this issue, we proposed a novel intelligent speed adaptation (ISA) system for visibility, called V-ISA (intelligent speed adaptation for visibility). It estimates the real-time safe speed limits based on the prevailing sight conditions. V-ISA comes with three variants with specific feedback modalities (1) visual and (2) auditory information, and (3) direct intervention to assume control over the vehicle speed. Here, we investigated the efficiency of each of the three V-ISA variants on driving speed choice and lateral behavioural response along road curves with limited and unsafe available sight distances, using a driving simulator. We also considered curve road geometry (curve direction: rightward vs. leftward). Sixty active drivers were recruited for the study. While half of them (experimental group) tested the three V-ISA variants (and a V-ISA off condition), the other half always drove with the V-ISA off (validation group). We used a linear mixed-effect model to evaluate the influence of V-ISA on driver behaviour. All V-ISA variants were efficient at reducing speeds at entrance points, with no discernible negative impact on driver lateral behaviour. On rightward curves, the V-ISA intervening variant appeared to be the most effective at adapting to sight limitations. Results of the current study implies that V-ISA might assist drivers to adjust their operating speed as per prevailing sight conditions and, consequently, establishes safer driving conditions. [ABSTRACT FROM AUTHOR]

Line-of-sight (LOS) sensors developed in newer vehicles have the potential to help avoid crash and near-crash scenarios with advanced driving-assistance systems; furthermore, connected vehicle technologies (CVT) also have a promising role in advancing vehicle safety. This study used crash and near-crash events from the Second Strategic Highway Research Program Naturalistic Driving Study (SHRP2 NDS) to reconstruct crash events so that the applicable benefit of sensors in LOS systems and CVT can be compared. The benefits of CVT over LOS systems include additional reaction time before a predicted crash, as well as a lower deceleration value needed to prevent a crash. This work acts as a baseline effort to determine the potential safety benefits of CVT-enabled systems over LOS sensors alone. [ABSTRACT FROM AUTHOR]

Many modern automated vehicle sensor systems use light detection and ranging (LiDAR) sensors. The prevailing technology is scanning LiDAR, where a collimated laser beam illuminates objects sequentially point-by-point to capture 3D range data. In current systems, the point clouds from the LiDAR sensors are mainly used for object detection. To estimate the velocity of an object of interest (OoI) in the point cloud, the tracking of the object or sensor data fusion is needed. Scanning LiDAR sensors show the motion distortion effect, which occurs when objects have a relative velocity to the sensor. Often, this effect is filtered, by using sensor data fusion, to use an undistorted point cloud for object detection. In this study, we developed a method using an artificial neural network to estimate an object's velocity and direction of motion in the sensor's field of view (FoV) based on the motion distortion effect without any sensor data fusion. This network was trained and evaluated with a synthetic dataset featuring the motion distortion effect. With the method presented in this paper, one can estimate the velocity and direction of an OoI that moves independently from the sensor from a single point cloud using only one single sensor. The method achieves a root mean squared error (RMSE) of 0.1187 m s−1 and a two-sigma confidence interval of [ − 0.0008 m s−1, 0.0017 m s−1] for the axis-wise estimation of an object's relative velocity, and an RMSE of 0.0815 m s−1 and a two-sigma confidence interval of [ 0.0138 m s−1, 0.0170 m s−1] for the estimation of the resultant velocity. The extracted velocity information (4D-LiDAR) is available for motion prediction and object tracking and can lead to more reliable velocity data due to more redundancy for sensor data fusion. [ABSTRACT FROM AUTHOR]

Abstract Cooperative adaptive cruise control (CACC) realizes efficient, intelligent control of vehicle acceleration, deceleration, and steering, through inter‐vehicle communication and cooperative control. However, the close combination of the platoon makes it difficult for other vehicles to cut‐in, which can lead to severe traffic jams on certain sections of the road. The control effect of the CACC depends on the platoon penetration rate, which is the percentage of connected and autonomous vehicles (CAVs) in the total number of platoon members. There is no quantitative control method for different penetration rates, and it is difficult to quantify the impact of CACC vehicles on traffic. Therefore, this paper proposes an innovative CACC control method based on deep reinforcement learning (DRL). First, the altruism control and the quantitative control of the car‐following strategy are realized by the virtual car‐following distance method to reduce the exclusivity of the CACC platoon or improve the road utilization efficiency. Second, a more appropriate platoon reward function and collision avoidance method are proposed. Finally, the Car Learning to Act (CARLA) simulator is used. The obtained results confirm that the CACC control of CAVs based on DRL can absorb speed oscillation and improve fuel economy.

Abstract Distracted driving is the leading cause of road traffic accidents. It is essential to monitor the driver's status to avoid traffic accidents caused by distracted driving. Current research on detecting distracting behaviours focuses on analysing image features using convolutional neural networks (CNNs). However, the generalisation ability of the current distracted driving models is limited. This paper aims to improve the generalisation ability of distracted driving models that are affected by factors such as the driver himself, the background, the monitoring angle, and so on. A new driver distraction detection method, which is referred to as multi‐scale domain adaptation network (MSDAN), was proposed to improve model adaptability. The method consists of three stages: first, multi‐scale convolution was introduced to build a new backbone to accommodate better the valuable feature of the target on different scales. Secondly, the authors designed the domain adaptation network to improve the model's adaptability to the difference in data sources through adversarial training. Finally, dropout is added to the fully connected layer to increase the model's generalisation ability. The comparison results on the large‐scale driver distraction detection dataset show that the authors’ method can accurately detect driver distraction and has good generalisation performance, with an accuracy improvement in the cross‐driver and cross‐dataset experiments.

Artificial intelligence (AI) will change society in so many different aspects that it is not comprehensible to grasp yet. It will be on the same magnitude as the Industrial Revolution (or most likely much larger). AI carries threats and great opportunities. Given the unknown unknowns of AI, we can spot headlines every day with either alarming news or the opposite. AI plays a major role for enabling autonomous vehicles but also for the development of more sophisticated and efficient advanced driver assistance systems (ADASs) by supporting, for example, rapid object detection and classification. Since the mid-20th century, AI has been a research topic, but not until recently has it gained momentum due to, for example, the performance increase in ubiquitous hardware. It is computationally expensive to execute AI algorithms. There are broadly speaking three “types” of AI discussed today, machine learning (ML), deep learning, and generative AI (GenAI), where GenAI is a specialized case of deep learning, which in turn is a subset of ML. The common denominator for all AI is training data, and data fit for purpose are pivotal for the development of AI algorithms. Garbage in will result in garbage out. [ABSTRACT FROM AUTHOR]

When a ground vehicle runs at high speeds, even a slight excess in the wheel steering angle can immediately cause the vehicle to slide sideways and lose control. In this study, we propose an active safety control system designed to address emergency situations where the driver applies excessive steering input and the vehicle speed varies significantly during control. The system combines the direct yaw moment (DYM) method with a steering saturation scheme that prevents excessive driver steering input from adversely influencing the front-wheel steering. Consequently, the control system allows the DYM to focus more on other stabilization tasks and maintain tire/road friction within its workable linear range. The implementation relies on a reference steering angle and a reference vehicle state, derived from a linear vehicle model considering tire/road friction limitations. When the driver’s steering angle and the system state deviate from these reference values, the control system intervenes by applying both the steering saturation scheme and DYM method. This ensures the front-wheel steering angle and system state remain close to the reference values. The control strategy is developed using the polytopic Linear Parameter Varying (LPV) technique and Linear Matrix Inequality (LMI) to account for the changes in vehicle speed. It is further enhanced with an input saturation technique based on a high-gain approach, which improves control utilization and system response during emergency situations. The advantages of the proposed control strategy are demonstrated through simulation results.

Abstract This paper develops a pedestrian potentially dangerous behaviour prediction method based on attention‐long‐short‐term memory (Attention‐LSTM) architecture to predict pedestrian trajectory and intention for the unexpected pedestrian crossing accident avoidance. To extract the road scene information for short periods of time, and improve the accuracy of subsequent intention inference and trajectory prediction, the panoramic segmentation is used to extrapolate pedestrian instances and segment areas of the environment. Next, an encoder–decoder framework based on Attention‐LSTM model is proposed to infer a pedestrian's intention to run or walk out into oncoming traffic straight and to predict the future trajectory. The proposed network involves two parts: temporal feature encoder and multi‐task decoder. The temporal feature encoder is mainly used to selectively emphasize the temporal features using attention mechanism, and then LSTM is employed for its encoding. In the multi‐task decoder, a multi‐head self‐attention mechanism and LSTM are used to forecast the pedestrians’ intention and future trajectory, respectively. Extensive experiments on pedestrian intention estimation (PIE) datasets demonstrate that the authors’ proposed approach surpasses prior studies in terms of prediction accuracy in trajectory and intention. This study can not only effectively avoid serious accidents caused by illegal road crossing but also achieve early warning for collision avoidance.

Recognizing speed limit information is crucial for advanced driver assistance systems (ADAS) as it directly affects the safety planning and decision-making process of intelligent driving systems. However, traditional image recognition-based solutions confront inherent restrictions and precision issues due to uncontrolled external factors. This paper endeavors to present a novel, data-driven solution for speed limit information recognition that leverages the stability and maturity of data-driven technologies, overcoming these challenges. We introduce Pareto-GBDTMO, a cutting-edge method that synergistically blends Gradient Boosting Decision Trees for Multiple Output (GBDT-MO) and Fast Pareto Feature Selection (FPFS). This integration is instrumental in discerning salient features to direct and expedite the learning process of GBDT-MO. When coupled with Bayesian optimization, the feature set undergoes dynamic updates at each boosting iteration, allowing GBDT-MO to concentrate on the most prominent features. This adaptive, relevance-guided feature space regularization mechanism enhances the efficiency and precision of speed limit recognition. Fujian Province highway electronic toll collection (ETC) data is used for further validation, and the experimental results emphasize the effectiveness of our model, with an high accuracy of 97%, a low loss rate of 0.7%, and minimal latency. These findings affirm the feasibility and scientific validity of our data-driven approach, offering a reliable and redundant solution for speed limit information recognition in ADAS. This study not only contributes to the practical application of ADAS but also lays the groundwork for future large-scale lane-level data research

Detecting drowsiness among drivers is critical for ensuring road safety and preventing accidents caused by drowsy or fatigued driving. Research on yawn detection among drivers has great significance in improving traffic safety. Although various studies have taken place where deep learning-based approaches are being proposed, there is still room for improvement to develop better and more accurate drowsiness detection systems using behavioral features such as mouth and eye movement. This study proposes a deep neural network architecture for drowsiness detection employing a convolutional neural network (CNN) for driver drowsiness detection. Experiments involve using the DLIB library to locate key facial points to calculate the mouth aspect ratio (MAR). To compensate for the small dataset, data augmentation is performed for the 'yawning' and 'no_yawning' classes. Models are trained and tested involving the original and augmented dataset to analyze the impact on model performance. Experimental results demonstrate that the proposed CNN model achieves an average accuracy of 96.69%. Performance comparison with existing state-of-the-art approaches shows better performance of the proposed model. [ABSTRACT FROM AUTHOR]