Browse Topic: Crash prevention

Items (133)
Find
Privacy-Preserving UAV-Based Traffic Data Acquisition Platform2026-99-05957/10/2026
This paper puts forward a Privacy-Preserving UAV-Based Traffic Data Acquisition Platform to address 1) privacy leakage, 2) limited scenario coverage, and 3) low traffic data utilization efficiency in urban traffic monitoring environments. Our system integrates three innovations: 1) Dynamic Privacy Masking (DPM) and Dual-Track acquisition (DTC), which hides sensitive information (e.g., faces, license plates or LPL) in real-time while preserving critical traffic data (e.g., vehicle density, speed), 2) traffic data Localization (DL) and Privacy-Enhanced Federated Learning (FEFL), enabling cross-regional collaboration without raw traffic data sharing by perturbing neural network updates with differential privacy (DP), and 3) Ground-Air Collaboration (GAC) and VPF (VPF), combining UAVs with ground sensors and digital twins (DTs) to cover blind spots (e.g., tunnels, extreme weather). Experimented on UA-DETRAC and CitySim traffic data-sets, the platform achieves 92% privacy compliance (GDPR/PIPL), 87.5% mAP accuracy, and 85% road network coverage, outperforming other methods (e.g., FedUAV, static blurring). It supports applications such as traffic flow optimization, accident prevention, and regulatory alignment.
Zhang, Shilin, Yan, Ming
The Influence of Spatial and Aspatial Factors on the Safety Driving Behavior of Motorcycle Couriers in the Gig-Economy Era: A Case Study of East Jakarta, Indonesia09-14-02-00057/3/2026
This study aims to analyze the impact of spatial and aspatial factors on the safety driving behavior of motorcycle couriers in East Jakarta within the context of the gig economy. Both factors are integrated to clarify how spatial conditions and individual characteristics jointly shape couriers’ safety driving behavior. The Partial Least Squares Structural Equation Modeling (PLS-SEM) method was employed to examine the relationship between spatial and aspatial factors on safety driving behavior. Data were collected through questionnaires from 253 motorcycle couriers operating in three subdistricts in East Jakarta, namely Cakung, Pasar Rebo, and Pulo Gadung. The results show that safety driving behavior is significantly influenced by aspatial factors, particularly socioeconomic characteristics and personality traits. In contrast, spatial factors such as road conditions and daily activity patterns do not directly influence safety driving behavior, but exert indirect effects through the couriers’ personality traits.
Wahyuddin, Yasser, Sitorus, Paldibo Alfriramson, Putri, Kharunia, Maharani, Garnierita
Development of Advanced Rider Assistance System (ARAS) for Electric Motorcycles: A Level-1 Collision Avoidance and Emergency Braking Prototype Using Ultrasonic Sensors2026-28-01182/12/2026
This paper presents the design, development, and validation of an Advanced Rider Assistance System (ARAS) tailored for electric motorcycles, with a specific focus on a Level-1 collision-avoidance and emergency-braking prototype employing ultrasonic sensing. The study is motivated by the disproportionately high accident exposure of two-wheeler riders and the slow adoption of ARAS technologies relative to the well-established Advanced Driver Assistance Systems (ADAS) in passenger vehicles. The proposed system utilizes front and rear ultrasonic sensors operating at 40 kHz, offering a measurement range of 2 cm to 4 m with ±1% accuracy, and maintaining reliable performance at motorcycle lean angles of up to 30°. Sensor data are processed using an STM32-series microcontroller running a real-time collision-risk estimation algorithm based on obstacle distance and relative velocity. A configurable safety threshold (typically 3 m) initiates a hierarchical warning strategy comprising visual indicators, acoustic alerts, and haptic cues. If the rider fails to respond within 300 ms, the system autonomously actuates emergency braking through a solenoid-based mechanism capable of modulating deceleration up to 0.6 g to maintain vehicle stability and avoid wheel lock. The ARAS prototype was developed through a structured workflow that included simulation of accident scenarios in IPG Motorcycle Maker, systematic component evaluation, and iterative hardware prototyping. Both simulation and on-road evaluations conducted at operating speeds of 10–40 km/h and various lean angles demonstrated consistent obstacle detection, prompt warning activation, and reliable emergency-braking performance. The system achieved a 92% reduction in simulated rear-end collisions and an average end-to-end response time of 180 ms. The modular system architecture further enables integration of additional sensing modalities and communication interfaces, providing a viable pathway toward higher levels of rider assistance. Overall, the study confirms the technical feasibility and safety benefits of a low-cost ultrasonic-based ARAS for electric motorcycles and establishes a strong foundation for broader deployment and future advancements in two-wheeler safety systems..
Deepan Kumar, Sadhasivam, Karu, Ragupathy, Karthick, K N, R, Vishnu Ramesh Kumar, Kumar, V, Manojkumar, R, M, Karthick, M, Rishab
Safer Trucks for India: Smart Tech Fusion2026-26-00311/16/2026
Commercial vehicle sector (especially trucks) has a major role in economic growth of a nation. With improving infrastructure, increasing number of trucks on roads, accidents are also increasing. As per RASSI (Road Accident Sampling System India) FY2016-23 database, commercial vehicles are involved in 42% of total accidents on Indian roads. Involvement of trucks (N2 & N3) is over 25% of total accidents. Amongst all accident scenarios of N2 &N3, frontal impacts are the most frequent (26%) and causing severe occupant injuries. Today, truck safety development for frontal impact is based on passive safety regulations (viz. front pendulum – AIS029) and basic safety features like seatbelts. In any truck accident, it is challenging rather impossible to manage comprehensive safety only with passive safety systems due to size and weight. Accident prevention becomes imperative in truck safety development due to extremely high energy involved in front impact scenarios. The paper presents a unique safety development approach (for frontal impact safety development for N2 and N3 trucks) which enables smart synthesis of active and passive safety systems to comprehensively address real world safety. Four major areas are identified for truck safety development viz. structural crashworthiness, compatibility, occupant safety and ADAS (Advanced Driver Assistance System). The innovation lies in smart mix of these areas during product safety development. The study presents the safety development of light commercial vehicle (truck) with this approach. Structural crashworthiness & occupant safety are developed with extensive number of CAE simulations. Design is physically validated with frontal impact test. In addition, extensive mileage accumulation is generated across Indian roads to validate ADAS system performance.
Joshi, Kedar Shrikant, Gadekar, Ganesh, Date, Atul, Koralla, Sivaprasad
DISRUPT - Decentralized Intelligent System for Road User Prediction and Tracking2025-01-02947/2/2025
We present DISRUPT, a research project to develop a cooperative traffic perception and prediction system based on networked infrastructure and vehicle sensors. Decentralized tracking and prediction algorithms are used to estimate the dynamic state of road users and predict their state in the near future. Compared to centralized approaches, which currently dominate traffic perception, decentralized algorithms offer advantages such as greater flexibility, robustness and scalability. Mobile sensor boxes are used as infrastructure sensors and the locally calculated state estimates are communicated in such a way that they can augment local estimates from other sensor boxes and/or vehicles. In addition, the information is transferred to a cloud that collects the local estimates and provides traffic visualization functionalities. The prediction module then calculates the future dynamic state based on neurocognitive behavior models and a measure of a road user's risk of being involved in dangerous situations. Based on this measure, alerts are generated and transmitted to road users equipped with an accident prevention app. An important component of DISRUPT is the development of a digital twin for testing and optimizing the overall system and its individual components. The main feature of the digital twin is the simulation of a photorealistic virtual copy of the test field environment. This enables the simulation of radar, infrared and conventional visible light cameras, which are combined with simulated data transmission delays to replicate the real system as accurately as possible. The plug-and-play design of the digital twin, together with a toolset for running and analyzing numerous simulations, enables efficient and thorough testing of the tracking and prediction algorithms. In particular, the digital twin enables the generation of hazard scenarios that are very unlikely to be observed in everyday traffic.
Beutenmüller, Frank, Brostek, Lukas, Doberstein, Christian, Han, Longfei, Kefferpütz, Klaus, Obstbaum, Martin, Pawlowski, Antonia, Rössert, Christian, Sas-Brunschier, Lucas, Schön, Thilo, Sichermann, Jörg
Early Detection of the Loss of Control of Motorsport Vehicles10-09-03-00296/30/2025
The article investigates how to detect as quickly as possible whether the driver will lose control of a vehicle, after a disturbance has occurred. Typical disturbances refer to wind gusts, obstacle avoidance, a sudden steer, traversing a pothole, a kick by another vehicle, and so on. The driver may be either human or non-human. Focus will be devoted to human drivers, but the extension to automated or autonomous cars is straightforward. Since the dynamic behavior of vehicle and driver is described by a saddle-type limit cycle, a proper theory is developed to use the limit cycle as a reference trajectory to forecast the loss of control. The Floquet theory has been used to compute a scalar index to forecast stable or unstable motion. The scalar index, named degree of stability (DoS), is computed very early, in the best case, in a few milliseconds after the disturbance has ended. Investigations have been performed at a dynamic driving simulator. A 14 DoF vehicle model, virtually driven by a real human driver, was employed. A number of evasive maneuvers have been examined, both for understeer and oversteer vehicles. The early detection of the loss of control is possible. The sensing of the loss of control could be enhanced with respect to a classical ESP, although a more in-depth investigation is needed. Some issues referring to the robustness of the computation of the DoS are still to be investigated. Nonetheless the DoS seems already applicable for motorsport vehicle and drivers.
Della Rossa, Fabio, Fontana, Matteo, Giacintucci, Samuele, Gobbi, Massimiliano, Mastinu, Giampiero, Previati, Giorgio
A Dynamic Hydroplaning Study: A Fluid–Structure Interaction Model for Road Vehicle Tire Analysis15-18-03-00156/24/2025
Hydroplaning contributes to approximately 20% of traffic accidents during adverse weather conditions, with factors such as velocity, water film thickness, tire inflation, and vehicle weight playing significant roles. This study aims to simulate the hydroplaning phenomenon using a fluid–structure interaction model based on the coupled Eulerian–Lagrangian (CEL) capabilities of ABAQUS. Results reveal that vehicle linear velocity is a key determinant of hydroplaning risk, with a positive correlation observed. The findings suggest maintaining speeds under 50 km/h to mitigate hydroplaning risk, contingent on well-maintained, properly inflated tires. Multiple linear regression analysis further demonstrates correlations among velocity, tire inflation, quarter vehicle load, and water film thickness in predicting the reaction force between the tire and roadway. The proposed scheme provides a predictive mechanism for hydroplaning risk under varying conditions, offering valuable insights into prevention strategies. The proposed scheme offers a valuable predictive mechanism for understanding and mitigating hydroplaning risk by analyzing key environmental and vehicle parameters. It identifies the critical factors influencing hydroplaning, including velocity, tire inflation, water film thickness, and vehicle load, while offering actionable insights to reduce risk. By employing advanced simulation techniques, specifically ABAQUS with CEL capabilities, the model provides a realistic and accurate representation of the hydroplaning phenomenon. Furthermore, the correlation analysis offers a comprehensive understanding of the relationship between multiple variables, enabling risk assessment under varying conditions. This approach not only highlights the underlying physics of hydroplaning but also supports evidence-based strategies for risk reduction and improved vehicle safety.
Aboelsaoud, Mostafa, Taha, Ahmed Abdelsalam, Abo Elazm, Mohamed, Elgamal, Hassan Anwar
Analysis and Evaluation of Roll Safety Thresholds of Tractor Semi-Trailer Vehicles during Turning Maneuvers Based on the Yaw Rate of Vehicle Bodies and Roll Safety Factor02-18-04-00246/5/2025
This article aims to analyze and evaluate the roll safety thresholds (RSTs) and roll safety zones of tractor semi-trailer vehicles during turning maneuvers, using the roll safety factor (RSF) and yaw rate of the vehicle bodies. To achieve this, a full dynamics model is established using the multibody system method. This model is then used to survey and evaluate the vehicle’s motion state, using ramp steer maneuver (RSM) steering rules. In each survey case, the maximum values of RSF and yaw rate of vehicle bodies are synthesized in 3D data, with an initial velocity range of 40 km/h to 80 km/h and a magnitude of steering wheel angle range of 12.5° to 300°. These 3D data are used to determine the proposed values of RSF, which can be used as examples to set the threshold values of the yaw rate of vehicle bodies and roll safety zones. At a velocity of 60 km/h, the dynamic rollover threshold for proposed roll safety factor (RSFprop) is equal to 1, with corresponding values of 15.718°/s and 14.962°/s. Similarly, the warning threshold for RSFprop is equal to 0.6, with values of 9.514°/s and 9.404°/s, and for RSFprop equal to 0.7, the values are 10.705°/s and 10.625°/s. The control threshold for a vehicle velocity of 60 km/h and RSFprop equal to 0.9 is calculated as 13.588°/s and 13.339°/s. These results can be used as a basis for developing early warning and control systems for various vehicle operating modes.
Hung, Ta Tuan
LightNetDehaze: Lightweight and Unsupervised Single-Image Dense Haze Removal12-09-01-00055/26/2025
Image dehazing techniques can play a vital role in object detection, surveillance, and accident prevention, especially in scenarios where visibility is compromised because of light scattering by atmospheric particles. To obtain a high-quality image or as an initial step in processing, it’s crucial to restore the scene’s information from a single image, given that this is an ill-posed inverse problem. The present approach utilized an unsupervised learning approach to predict the transmission map from a hazy image and used YOLOv8n to detect the car from a clear recovered image. The dehazing model utilized a lightweight parallel channel architecture to extract features from the input image and estimate the transmission map. The clear image is recovered using an atmospheric scattering model and given to the YOLOv8n for car detection. By incorporating dark channel prior loss during training, the model eliminates the need for a paired dataset. The proposed dehazing model with fewer parameters speeds up the dehazing process, which can detect the objects in less response time. The network follows unsupervised learning, which eliminates the need of ground truth image or transmission map of a clear image. The proposed method tried to solve the issue of high computational complexity and long latency when used as a preprocessing stage in computer vision applications. The proposed network ranks first in terms of parameters and FLOPs, which are lower by scale 102 and 103, respectively, compared to the method ranked second. The results highlight the effectiveness of the proposed method compared to other methods and ranked first in number of car detections using YOLOv8n. The inference time to dehaze the image is comparable to the method ranked first and 66% lower than the third rank.
Dave, Chintan, Patel, Hetal, Kumar, Ahlad
Predicting Real-time Crash Risks during Hurricane Evacuation Using Connected Vehicle Data09-13-01-00064/12/2025
Hurricane evacuations generate high traffic demand with increased crash risk. To mitigate such risk, transportation agencies can adopt high-resolution vehicle data to predict real-time crash risks. Previous crash risk prediction models mainly used limited infrastructure sensor data without covering many road segments. In this article, we present methods to determine potential crash risks during hurricane evacuation from an emerging alternative data source known as connected vehicle data that contain vehicle speed and acceleration information collected at a high frequency (mean = 14.32, standard deviation = 6.82 s). The dataset was extracted from a database of connected vehicle data for the evacuation period of Hurricane Ida on Interstate-10 in Louisiana. Five machine learning models were trained considering weather features and different traffic characteristics extracted from the connected vehicle data. The results indicate that the Gaussian process boosting and extreme gradient boosting models outperform (recall = 0.91) other models. Such real-time crash prediction models, leveraging connected vehicle data, could enable transportation agencies to implement proactive countermeasures, such as dynamic speed limit or lane management, during emergency evacuations, thereby enhancing road safety.
Syed, Zaheen E Muktadi, Hasan, Samiul
A Low-Entry Bus Body Structure Performance under the Conditions of the New Car Assessment Program Frontal Crash Tests09-13-02-00074/9/2025
The New Car Assessment Program (e.g., US NCAP and EuroNCAP) frontal crash tests are an essential part of vehicle safety evaluations, which are mandatory for the certification of civil means of transport prior to normal road exploitation. The presented research is focused on the behavior of a tubular low-entry bus frame during a frontal impact test at speeds of 32 and 56 km/h, perpendicular to a rigid wall surface. The deformation zones in the bus front and roof parts were estimated using Ansys LS-DYNA and considered such factors as the additional mass (1630 kg) of electric batteries following the replacement of a diesel engine with an electric one. This caused stabilization of the electric bus body along the transverse axis, with deviations decreased by 19.9%. Speed drop from 56 to 32 km/h showed a reduction of the front window sill deformations from 172 to 132 mm, and provided a twofold margin (159.4 m/s2) according to the 30g ThAC criterion of R80. This leads to the conclusion about the recommendation to reduce the regulatory NCAP speed for city buses to ensure a sufficient safety margin, especially considering the city speed limit of 40–50 km/h. The novelty consists in the application of the NCAP regulations, which is not typical for buses, with the formation of conclusions about the higher danger for passengers compared to UNECE R29 due to the much deeper penetration of plastic deformations into the bus body (up to the rear axle). The obtained results can influence accident prevention strategies and the revision of the city bus certification approaches and regulations.
Holenko, Kostyantyn, Dykha, Aleksandr, Koda, Eugeniusz, Kernytskyy, Ivan, Royko, Yuriy, Horbay, Orest, Berezovetska, Oksana, Rys, Vasyl, Humeniuk, Ruslan, Berezovetskyi, Serhii, Chalecki, Marek
Tube-MPC Based Path Tracking for I Vehicle Considering Data-Driven Stable Region and External Disturbances2025-01-82624/1/2025
Path tracking is a key function of intelligent vehicles, which is the basis for the development and realization of advanced autonomous driving. However, the imprecision of the control model and external disturbances such as wind and sudden road conditions will affect the path tracking effect and even lead to accidents. This paper proposes an intelligent vehicle path tracking strategy based on Tube-MPC and data-driven stable region to enhance vehicle stability and path tracking performance in the presence of external interference. Using BP-NN combined with the state-of-the-art energy valley optimization algorithm, the five eigenvalues of the stable region of the vehicle β−β̇ phase plane are obtained, which are used as constraints for the Tube-MPC controller and converted into quadratic forms for easy calculation. In the calculation of Tube invariant sets, reachable sets are used instead of robust positive invariant sets to reduce the calculation. Simulation results demonstrates that the strategy is with superior performance than conventional MPC without considering stability under the condition with different road adhesion coefficients and lateral winds.
Zhang, Haosen, Li, Yihang, Wu, Guangqiang
Avoiding the Crash: A Vision-Language Model Evaluation of Critical Traffic Scenarios2025-01-82134/1/2025
Autonomous Vehicles (AVs) have transformed transportation by reducing human error and enhancing traffic efficiency, driven by deep neural network (DNN) models that power image classification and object detection. However, to maintain optimal performance, these models require periodic re-training; failure to do so can result in malfunctions that may lead to accidents. Recently, Vision-Language Models (VLMs), such as LLaVA-7B and MoE-LLaVA, have emerged as powerful alternatives, capable of correlating visual and textual data with a high degree of accuracy. These models’ robustness and ability to generalize across diverse environments make them especially suited to analyzing complex driving scenarios like crashes. To evaluate the decision-making capabilities of these models across common crash scenarios, a set of real-world crash incident videos was collected. By decomposing these videos into frame-by-frame images, we task the VLMs to determine the appropriate driving action at each frame: accelerate, brake, turn left, turn right, or maintain the current course. For each frame, three sets of outputs are analyzed: the actual action executed in the video, the action a human driver would likely take to avoid a crash, and the action the VLM predicts as optimal to avoid a crash. To measure and compare the effectiveness of the VLMs, we introduce a metric called Crash Prevention Efficiency (CPE) which evaluates the model’s performance in detecting crash scenarios and taking appropriate actions to avoid them. CPE assesses how well a VLM can respond to potential crashes by analyzing both the timing of the detection and the proximity to a predefined point in the crash sequence. Our findings reveal that VLMs demonstrate a high level of consistency in decision-making, with LLaVA-7B and MoE-LLaVA models identifying potential crash scenarios 1.13 to 1.33 seconds earlier than humans, respectively. This highlights their potential role in autonomous driving systems (ADS), supporting both real-time decision-making for human drivers and fully autonomous operations.1
Fernandez, David, MohajerAnsari, Pedram, Salarpour, Amir, Pesé, Mert D.
Compensation Algorithm Development in Platform for Pedestrian ADAS Testing2025-01-80664/1/2025
Test procedures such as EuroNCAP, NHTSA’s FMVSS 127, and UNECE 152 all require specific pedestrian to vehicle overlaps. These overlap variations allow the vehicle differing amounts of time to respond to the pedestrian’s presence. In this work, a compensation algorithm was developed to be used with the STRIDE robot for Pedestrian Automatic Emergency Braking tests. The compensation algorithm uses information about the robot and vehicle speeds and positions determine whether the robot needs to move faster or slower in order to properly overlap the vehicle. In addition to presenting the algorithm, tests were performed which demonstrate the function of the compensation algorithm. These tests include repeatability, overlap testing, vehicle speed variation, and abort logic tests. For these tests of the robot involving vehicle data, a method of replaying vehicle data via UDP was used to provide the same vehicle stimulus to the robot during every trial without a robotic driver in the vehicle.
Bartholomew, Meredith, Nguyen, An, Helber, Nicholas, Heydinger, Gary
Daytime and Nighttime Performance Evaluations of the Automatic Emergency Braking and Forward Collision Warning Systems on a 2020 and 2022 Kia Telluride2025-01-86994/1/2025
Testing was conducted in daytime and nighttime conditions to evaluate the performance of the Automatic Emergency Braking and Forward Collision Warning systems present on both a 2020 and 2022 Kia Telluride. The 2022 Kia Telluride was tested during the day at speeds between 35 and 70 miles per hour, while the 2020 Kia Telluride was tested both during the day and at night at speeds between 35 and 60 miles per hour (mph). The daytime testing of both the 2020 and 2022 Kia Telluride utilized a foam stationary vehicle target. The nighttime testing of the 2020 Kia Telluride utilized a live 2006 Chevrolet Tahoe as the target with the brake lights on. Testing measured the Time to Collision (TTC) values of the visual/audible component of the Forward Collision Warning (FCW) that was presented to the driver. Further, testing also quantified the timing and magnitude of the two-phase response of the Automatic Emergency Braking (AEB) system. The results of both sets of testing add higher speed FCW and AEB testing scenarios to the database of publicly available tests for the Kia Telluride.
Harrington, Shawn, Patrick-Moline, Peyton, Nagarajan, Sundar Raman
An Evaluation of the Performance of the Automatic Emergency Braking and Forward Collision Warning System in a 2020 Jeep Grand Cherokee during Daytime and Nighttime Conditions2025-01-86984/1/2025
Testing was conducted to evaluate the performance of the 2020 Jeep Grand Cherokee’s Forward Collision Warning (FCW) and Automatic Emergency Braking (AEB) collision mitigation systems at speeds between 35 and 70 miles per hour (mph). Two different 2020 Jeep Grand Cherokee’s were utilized under varying testing conditions in order to evaluate the performance of their collision mitigation systems. A total of 40 tests were conducted: 29 tests were conducted during daytime and 11 tests were conducted at nighttime. Testing measured the Time to Collision (TTC) values of the visual/audible component of the Forward Collision Warning that was presented to the driver. In addition, the testing quantified the TTC response of the Automatic Emergency Braking (AEB) system including the timing and magnitude of the automatic braking response. The results of the testing add higher speed FCW and AEB testing scenarios to the database of publicly available tests for the 2020 Jeep Grand Cherokee.
Harrington, Shawn, Lieber, Victoria, Nagarajan, Sundar Raman
Advanced Driver Assistance Systems (ADAS): Assessing the Efficacy of Non-Impact Testing for Evaluating the Performance of Frontal Collision Mitigation Technology2025-01-80564/1/2025
As Automatic Emergency Braking (AEB) systems become standard equipment in more light duty vehicles, the ability to evaluate these systems efficiently is becoming critical to regulatory agencies and manufacturers. A key driver of the practicality of evaluating these systems’ performance is the potential collision between the subject vehicle and test target. AEB performance can depend on vehicle-to-vehicle closing speeds, crash scenarios, and nuanced differences between various situational and environmental factors. Consequently, high speed impacts that may occur while evaluating the performance of an AEB system, as a result of partial or incomplete mitigation by an AEB activation, can cause significant damage to both the test vehicle and equipment, which may be impractical. For tests in which impact with the test target is not acceptable, or as a means of increasing test count, an alternative test termination methodology may be used. One such method constitutes the application of a late steering maneuver by the driver to avoid the target prior to a potential collision. In this study, a test series was performed with and without late steering input to determine if this alternative AEB evaluation methodology can be used to accurately predict the degree to which an AEB system performs. The results were compared to non-swerve tests to assess whether non-impact testing can be used to determine the extent to which AEB system response would have mitigated a collision. The findings indicate that there are significant limitations to the accuracy of predictions made with this approach.
Kuykendal, Michelle, Easter, Casey, Koszegi, Giacomo, Alexander, Ross, Paradiso, Marc, Scally, Sean
Research on Dangerous Driving Behavior Recognition Algorithm Based on LSTM2025-01-82044/1/2025
To address the issue of high accident rates in road traffic due to dangerous driving behaviors, this paper proposes a recognition algorithm for dangerous driving behaviors based on Long Short-Term Memory (LSTM) networks. Compared with traditional methods, this algorithm innovatively integrates high-frequency trajectory data, historical accident data, weather data, and features of the road network to accurately extract key temporal features that influence driving behavior. By modeling the behavioral data of high-accident-prone road sections, a comprehensive risk factor is consistent with historical accident-related driving conditions, and assess risks of current driving state. The study indicates that the model, in the conditions of movement track, weather, road network and conditions with other features, can accurately predict the consistent driving states in current and historical with accidents, to achieve an accuracy rate of 85% and F1 score of 0.82. It means the model can effectively detect the consistency of driving states in current and hazardous driving behaviors in historical, such as aggressive acceleration, abrupt deceleration, and speeding. It contributes to provide timely risk warnings for drivers, effectively reducing the occurrence of traffic accidents, and adherence to driving safety.
Huang, Yinuo, Zhang, Miaomiao, Xue, Ming, Jin, Xin
Longitudinal Analysis of Forward Blind Zone Changes in Popular Vehicle Models (1997–2023)09-13-01-00053/26/2025
Background. In 2022, vulnerable road user (VRU) deaths in the United States increased to their highest level in more than 40 years. At the same time, increasing vehicle size and taller front ends may contribute to larger forward blind zones, but little is known about the role that visual occlusion may play in this trend. Goal. Researchers measured the blind zones of six top-selling light-duty vehicle models (one pickup truck, three SUVs, and two passenger cars) across multiple redesign cycles (1997–2023) to determine whether the blind zones were getting larger. Method. To quantify the blind zones, the markerless method developed by the Insurance Institute for Highway Safety was used to calculate the occluded and visible areas at ground level in the forward 180° arc around the driver at ranges of 10 m and 20 m. Results. In the 10-m forward radius nearest the vehicle, outward visibility declined in all six vehicle models measured across time. The SUV models showed up to a 58% reduction in visibility within a 10 m radius. Other vehicles exhibited smaller (7%–19%) reductions. At longer distances (10 m–20 m), vehicles demonstrated both increases and decreases in visibility. Conclusion. The markerless method provides a straightforward and replicable assessment of driver visibility. The observed decrease in direct outward visibility near the vehicles points to the need for further study regarding this trend, including analysis of the repeatability and viability of the measurement technique.
Epstein, Alexander K., Brodeur, Alyssa, Drake, Juwon, Englin, Eric, Fisher, Donald L., Zoepf, Stephen, Mueller, Becky C., Bragg, Haden
Advancing Road Sensing Technology: Functional Observer for Vehicle Road Profile Estimation12-08-04-00383/24/2025
Having an in-depth comprehension of the variables that impact traffic is essential for guaranteeing the safety of all drivers and their automobiles. This means avoiding multiple types of accidents, particularly rollover accidents, that may have the capacity of causing terrible repercussions. The non-measured factors in the system state can be estimated employing a vehicle model incorporating an unknown input functional observer, this gives an accurate estimation of the unknown inputs such as the road profile. The goal of the proposed functional observer design constraints is to reduce the error of estimation converging to a value of zero, which results in an improved calculation of the observer parameters. This is accomplished by resolving linear matrix inequalities (LMIs) and employing Lyapunov–Krasovskii stability theory with convergence conditions. A simulator that enables a precise evaluation of environmental factors and fluctuating road conditions was additionally utilized. This research makes an important contribution to road safety via the development of cutting-edge technologies for vehicle control and monitoring.
Saber, Mohamed, Ouahi, Mohamed, Naami, Ghali, El Akchioui, Nabil
Improving Safety of Autonomous Vehicles: A Verifiable Method for Graceful Degradation of Decision and Control Responsibilities12-08-02-00213/24/2025
Developing safe and reliable autonomous vehicles is crucial for addressing contemporary mobility challenges. While the goal of autonomous vehicle development is full autonomy, up to SAE Level 4 and beyond, human intervention remains necessary in critical or unfamiliar driving scenarios. This article introduces a method for gracefully degrading system functionality and seamlessly transferring decision-making and control between the autonomous system and a remote safety operator when needed. This transfer is enabled by an onboard dependability cage, which continuously monitors the vehicle’s performance during its operation. The cage communicates with a remote command control center, allowing for remote supervision and intervention by a safety driver. We assess this methodology in both lab and test field settings in a case study of last-mile parcel delivery logistics and discuss the insights and results obtained from these evaluations.
Aniculaesei, Adina, Aslam, Iqra, Zhang, Meng, Buragohain, Abhishek, Vorwald, Andreas, Rausch, Andreas
AVSC Best Practice for Automated Driving System-Dedicated Vehicle (ADS-DV) Immediate Post-Crash Behaviors and InteractionsAVSC-I-05-20253/4/2025
The behaviors and interactions of automated driving system-dedicated vehicles (ADS-DVs) following a crash can have continuing effects on road safety. As more ADS-DVs are commercially deployed and tested on public roads, it is important that systems are designed not only to prevent crashes, but to minimize the impact severity in situations where crashes cannot be avoided. Further injuries, vehicle and property damage, and secondary crashes can be mitigated based on appropriate ADS-DV behaviors and interactions following a crash. This best practice defines five general crash stages, focusing on the Immediate Post-Crash stage. It presents a framework with examples of ADS-DV behavior and interaction activities. It includes recommendations on decision making and evaluation of immediate post-crash ADS-DV behaviors. The document also highlights three dependencies that impact the behavior and interaction of an ADS-DV. The dependencies are paired with illustrative scenarios, followed by questions and answers to explain possible considerations for the ADS-DV post crash behaviors and interactions. Stakeholder understanding of ADS-DV behaviors and interactions immediately following a crash supports the AVSC goal of helping to ensure safer, more predictable SAE level 4 and level 5 autonomous vehicles (AVs) and building public trust.
Automated Vehicle Safety Consortium
Effects of Emergency Vehicle Warning Lighting System Characteristics on Driver Perception and Behavior09-13-01-00032/28/2025
Secondary crashes, including struck-by incidents are a leading cause of line-of-duty deaths among emergency responders, such as firefighters, law enforcement officers, and emergency medical service providers. The introduction of light-emitting diode (LED) sources and advanced lighting control systems provides a wide range of options for emergency lighting configurations. This study investigated the impact of lighting color, intensity, modulation, and flash rate on driver behavior while traversing a traffic incident scene at night. The impact of retroreflective chevron markings in combination with lighting configurations, as well as the measurement of “moth-to-flame” effects of emergency lighting on drivers was also investigated. This human factors study recruited volunteers to drive a closed course traffic incident scene, at night under various experimental conditions. The simulated traffic incident was designed to replicate a fire apparatus in the center-block position. The incident scene was complemented with a cone taper extending from the driver-side buffer to the edge of the roadway. The results indicate that higher-intensity lights were judged consistently as more glaring, but were only rated as marginally more visible. The rated visibility of the lights appears to be related to the perceived saturation of the color, while discomfort glare is related to the amount of short-wavelength spectral content. The results also suggest that the presence of highly reflective markings may decrease drivers’ ability to see first responders working adjacent to their vehicles.
Bullough, John D., Parr, Scott, Hiebner, Emily, Sblendorio, Alec
Toward a Methodology for Safety—Performance Trade-Off Analysis for Connected Automated Vehicles Supported by a Smart Infrastructure12-08-02-00202/24/2025
Introducing connectivity and collaboration promises to address some of the safety challenges for automated vehicles (AVs), especially in scenarios where occlusions and rule-violating road users pose safety risks and challenges in reconciling performance and safety. This requires establishing new collaborative systems with connected vehicles, off-board perception systems, and a communication network. However, adding connectivity and information sharing not only requires infrastructure investments but also an improved understanding of the design space, the involved trade-offs and new failure modes. We set out to improve the understanding of the relationships between the constituents of a collaborative system to investigate design parameters influencing safety properties and their performance trade-offs. To this end we propose a methodology comprising models, analysis methods, and a software tool for design space exploration regarding the potential for safety enhancements and requirements on off-board perception systems, the communication network, and AV tactical safety behavior. The methodology is instantiated as a concrete set of models and a tool, exercised through a case study involving intersection traffic conflicts. We show how the age of information and observation uncertainty affect the collaborative system design space and further discuss the generalization and other findings from both the methodology and case study development.
Fornaro, Gianfilippo, Törngren, Martin, Gaspar Sánchez, José Manuel
Quantifying Hydroplaning Risk: Insights from Water Film Thickness Measurements on Maryland Highways02-18-02-00102/19/2025
Wet pavement conditions during rainfall present significant challenges to traffic safety by reducing tire–road friction and increasing the risk of hydroplaning. During high-intensity rain events, the roadway pavement tends to accumulate water, forming a film that can have serious implications for vehicle control. As the longitudinal speed of the vehicle increases, a water wedge forms in front of the tire, leading to partial loss of contact with the road. At critical hydroplaning speed, a complete water layer forms between the tire and the road. Although less common, dynamic hydroplaning poses severe risks when high-intensity rainfall coincides with high vehicle traveling speed, leading to a complete loss of control over vehicle steering capabilities. This study advances hydroplaning research by integrating real-world data from the Road Weather Information System (RWIS) with an existing hydroplaning model. This approach provides more accurate hydroplaning risk assessments, emphasizing the importance of adapting predictive models to real-world conditions. Measurements of water film thickness from two Maryland locations over a year showed values of the water film heights up to 1.9 mm, with significant hydroplaning risk for vehicles with worn tires traveling at highway speeds. Using models such as Gengenbach and Gallaway, the study computes critical hydroplaning speeds, highlighting the importance of tire tread depth, inflation pressure, and pavement texture. Results indicate that the critical hydroplaning speed varies significantly based on these factors, emphasizing the need for safe driving practices during heavy rainfall. The findings underscore also the importance of developing new hydroplaning models in the context of future autonomous vehicles that needs robust algorithms for operating in wet conditions.
Vilsan, Alexandru, Sandu, Corina, Anghelache, Gabriel
Validating Human Factors-Based Classification Models for Safe and Efficient Mixed-Autonomy Interactions at Intersections Using a Virtual Simulator12-08-02-00192/12/2025
The introduction of autonomous vehicles (AVs) promises significant improvements to road safety and traffic congestion. However, mixed-autonomy traffic remains a major challenge as AVs are ill-suited to cooperate with human drivers in complex scenarios like intersection navigation. Specifically, human drivers use social cooperation and cues to navigate intersections while AVs rely on conservative driving behaviors that can lead to rear-end collisions, frustration from other road users, and inefficient travel. Using a virtual driving simulator, this study investigates the use of a human factors-informed cooperation model to reduce AV reliance on conservative driving behaviors. Four intersection scenarios, each involving a left-turning AV and a human driver proceeding straight, were designed to obfuscate the right-of-way. The classification models were trained to predict the future priority-taking behavior of the human driver. Results indicate that AVs employing the human factors-informed model were able to navigate the mixed-autonomy intersection scenarios significantly more efficiently without affecting safety or rider comfort when compared to a baseline, cautious AV. Overall, this research contributes to improved mixed-autonomy interactions and provides evidence for the importance of cooperation between AVs and human-driven vehicles.
Ziraldo, Erika, Oliver, Michele
Real-Time Driver Drowsiness Detection Using YOLOv8 with Whale Optimization Algorithm09-13-01-00012/10/2025
Driver fatigue and drowsiness portray an integral role in the frequency of road accidents. Putting in place policies intended to alert drivers is imperative for averting accidents and saving lives. This work aims to improve road safety by devising a real-time driver drowsiness detection system. To accomplish this, drowsiness is detected using YOLOv8 algorithm optimized with the whale optimization algorithm (WOA). Key facial cues such as eye closure and yawning frequency are monitored to analyze driving behavior by the suggested approach. YOLOv8 model optimized with WOA processes video streams in real time and sets off an alarm on the graphical user interface (GUI) dashboard based on the output. The proposed approach was investigated using two datasets namely UTA-RLDD and D3S. A 640 × 640 pixel image with a frame rate of 50 fps was used in the investigation. The mAP at 0.5 (mean average precision at 0.5 IoU (intersection over union) threshold) of drowsiness detection system using UTA-RLDD dataset is 85.4% and using D3S dataset is 84.3%. It was indicated by the obtained results that the WOA-optimized YOLOv8 model attains superior detection accuracy and quicker inference times in comparison to the contemporary methods. This analysis sets the stage for advancements in real-time drowsiness detection techniques for implementation in vehicle safety systems.
Nandal, Priyanka, Pahal, Sudesh, Sharma, Tripti, Omesh, Omesh
Enhanced Adaptive Cruise Control Strategies Considering Front and Rear Vehicles10-09-01-00071/29/2025
Adaptive cruise control (ACC) systems have increasingly become more robust in adapting to the motion of the preceding vehicle and providing safety and comfort to the driver. But conventional ACC hangs with a concern for rear-end safety in the presence of traffic or aggressive car maneuvers. It often leads to getting dangerously close to the vehicle behind in scenarios where there is less space and time for the rear vehicle to adjust. This research article develops an ACC approach that considers the rear vehicle in addition to the front vehicle, thereby ensuring safety with the rear vehicle without compromising the safety of the front vehicle. Two novel methodologies are devised to enhance the ACC system. The first approach involves utilizing fuzzy logic to associate the inputs with the throttle and brake based on the inference rules within a fuzzy logic controller overseeing both vehicles. The other utilizes a cascaded model predictive control (MPC) system framework that integrates a novel formulation based on vehicle kinematics to devise an optimal reference speed to maintain safe distance with both vehicles, which is fed to the lower level MPC that generates the corresponding throttle and brake values to track the reference speed while ensuring smooth speed transitions. Priority is given to the front vehicle in conflicting situations. Finally, the efficacy of the control strategies is validated using industry-standard simulation software Prescan by assessing the comparative performances of both the control strategies. The results of this study will provide valuable insights into the enhancement of ACC systems by improving the distance safety margin of the ACC-equipped vehicle with respect to the rear vehicle along with the front, ultimately contributing to better throughput of traffic and safer road mobilization.
Sharma, Vishrut, Sengupta, Somnath, Ghosh, Susenjit
Letter from the Guest Editors12-08-01-000112/10/2024
Letter from the Guest Editors
van Schijndel, Margriet, Sciarretta, Antonio, Op den Camp, Olaf, Krosse, Bastiaan
Design and Implementation of Advanced Safety Features in Two-Wheeled Vehicles for Accident Prevention and Anti-Theft System2024-01-522912/10/2024
Embarking on exploring the cutting-edge domain of smart bike innovations, this study focuses primarily on enhancing safety and security measures. Through meticulous development and implementation, it introduces seven pioneering features to curb accidents and thwart theft incidents. These transformative functionalities encompass a spectrum of aspects, including cautionary systems for side stand and helmet usage, advanced alcohol detection mechanisms, and robust anti-theft measures employing ID card and password protocols. Moreover, integrating speed control mechanisms and automated brake activation on encountering speed breakers further elevates the safety quotient of the smart bike. By harnessing a diverse array of sensors such as RF, REED, ultrasonic, and gas sensors, these features collectively pave the way for a paradigm shift in road safety standards. The report meticulously details the intricacies of design, execution, and cost estimation, underscoring the transformative impact of these innovations in bolstering road safety and safeguarding against theft incidents.
Mallieswaran, K., Agaramudhalvan, S., Nithya, R., Shuruti, R., Radhika, S.
Speed Control Assurance for SnowmobilesJ1222_202412 (Current)12/5/2024
This SAE Standard provides minimum requirements and performance criteria for devices to prevent runaway snowmobiles due to malfunction of the speed control system.
Snowmobile Technical Committee
LDFA: Lightweight Dynamic Feature Aggregation for Multi-Modal Fusion2024-01-700811/15/2024
Integrating 3D point cloud and image fusion into flying car detection systems is essential for enhancing both safety and operational efficiency. Accurate environmental mapping and obstacle detection enable flying cars to optimize flight paths, mitigate collision risks, and perform effectively in diverse and challenging conditions. The AutoAlignV2 paradigm recently introduced a learnable schema that unifies these data formats for 3D object detection. However, the computational expense of the dynamic attention alignment mechanism poses a significant challenge. To address this, we propose a Lightweight Cross-modal Feature Dynamic Aggregation Module, which utilizes a model-driven feature alignment strategy. This module dynamically realigns heterogeneous features and selectively emphasizes salient aspects within both point cloud and image datasets, enhancing the differentiation between objects and the background and improving detection accuracy. Additionally, we introduce the Lightweight Spatial-Reduction Attention (LSRA) layer to enhance the original attention mechanism. By employing spatial reduction and positional offset techniques, LSRA reduces computational complexity, accelerating the aggregation of cross-modal features while minimizing computational overhead. Furthermore, we implement a novel dropout scheme before extracting features from 2D images, enhancing the model's generalization capabilities and reducing computational costs. We present a new lightweight framework—Lightweight Dynamic Feature Aggregation for Multi-modal Fusion (LDFA)—designed specifically for the harmonious fusion of 3D point cloud data and 2D image-derived information. The LDFA framework achieves a meticulous balance between computational efficiency and enhanced perceptual capabilities. Extensive experimental evaluations on the nuScenes benchmark dataset confirm the efficacy and efficiency of the LDFA fusion strategy, demonstrating its potential to redefine the state-of-the-art in multimodal 3D object detection. Code will be available at https://github.com/zishenjiucai/LDFA.
Feng, Xiaoyu, Zhang, Renhang, Chu, Zheng, Wei, Lina, Bian, Chen, Duan, Linshuai
Boosting Algorithms for the Accident Severity Classification12-08-04-003010/17/2024
Background: Road accident severity estimation is a critical aspect of road safety analysis and traffic management. Accurate severity estimation contributes to the formulation of effective road safety policies. Knowledge of the potential consequences of certain behaviors or conditions can contribute to safer driving practices. Identifying patterns of high-severity accidents allows for targeted improvements in terms of overall road safety. Objective: This study focuses on analyzing road accidents by utilizing real data, i.e., US road accidents open database called “CRSS.” It employs advanced machine learning models such as boosting algorithms such as LGBM, XGBoost, and CatBoost to predict accident severity classification based on various parameters. The study also aims to contribute to road safety by providing predictive insights for stakeholders, functional safety engineering community, and policymakers using KABCO classification systems. The article includes sections covering theoretical methodology, data analysis, model development, evaluation, performance metrics, and implications for improving road safety measures by comparing the performance of different boosting algorithms on the CRSS dataset. This study aims to identify the most effective machine learning algorithm to integrate into our product line in the near future, enabling accurate prediction of both accident severity and occurrence. Results and Conclusions: This study addresses challenges in evaluating performance metrics for different severity classes within unbalanced datasets, emphasizing the impact of dominant classes like Class O (O = no apparent injury) on overall accuracy. The investigation reveals the limitations and conservatism associated with imbalanced data in boosting models, hinting at a potential ceiling in their performance around 80%. Comparative analysis of algorithms, including CatBoost, XGBoost, and LGBM, demonstrates comparable performance even in the case of applying KNN algorithm for pre-processing, based on various metrics, especially accuracy, F 1-score, ROC-AUC, and PR-AUC for all severity classes. XGBoost with KNN algorithm did not show any significant performance improvement compared to the XGBoost without KNN algorithm. The study includes performance metrics, such as F 1-score, CM upper triangle, ROC-AUC, and PR-AUC applied to an accident analysis case study. Future work directions involve extending the application of CatBoost, XGBoost, and other algorithms to diverse datasets, exploring the capabilities of deep neural networks, refining dataset preparation for accuracy improvement, and creating unified tools for hazard analysis and risk assessment.
Babaev, Islam, Mozolin, Igor, Garikapati, Divya
Research on an Automatic Emergency Braking Control Method Based on Coordinated Wheel Deflection09-12-03-001110/16/2024
Automatic emergency braking (AEB) systems play a crucial role in enhancing vehicular safety. Current research predominantly focuses on the longitudinal dynamics of vehicles, utilizing various control algorithms to improve braking effectiveness. However, there has been limited exploration into utilizing wheel deflection as a method to further enhance emergency braking performance. This study aims to contribute by proposing an advanced enhancement of the AEB system through coordinated wheel deflection strategies. In an emergency situation, when the speed of AEB-equipped vehicle drops to the set threshold due to wheel braking, the innovative control system will activate. The vehicle’s coaxial wheels will then execute a counter-deflection maneuver to maximize friction between the tires and the road surface. As a result, this approach reduces braking distance, thereby enhancing vehicle safety. The effectiveness of the proposed control algorithm is validated through combined simulation using CarSim and MATLAB/Simulink.
Lai, Fei, Xiao, Hao, Huang, Chaoqun
SAE TOMORROW TODAY: Improving Road Safety With ADAS and ADS134718/1/2024
Thousands of people die or are injured in automobile crashes each year. By leveraging in-vehicle advanced driver assistance systems (ADAS) and automated driving systems (ADS) more proactively, errors caused by human drivers can be reduced, resulting in less fatal and serious car crashes. To learn more about the future of road safety, we sat down with Seth Chalmers, Director of Traffic Engineering for Dibble, to discuss new NHTSA developments, the intersection of ADAS and autonomy, and his experience with Waymo in Phoenix.
Hineman, Marcie
Scenario-Based Helicopter Flight Simulation of Accident-Prone Vortex Ring State (VRS) EncountersF-0080-2024-11415/7/2024
From 2008 to 2021, 48 helicopter accidents have involved Vortex Ring State (VRS) encounters in the United States. For dangerous situations such as VRS encounters and recoveries, Scenario-Based Training (SBT) in flight simulators could supplement flight training, allowing pilots to practice in a safe environment. In this study, we created and tested a proof of concept for scenario-based training in flight simulators, dedicated to VRS-related accident prevention. The goal is to evaluate pilots' awareness, avoidance, detection, and recovery skills during VRS-inducing scenarios. Moreover, this study intends to provide a holistic analysis of VRS encounters and recoveries by examining the situations and factors leading to VRS-related crashes, and to lay out the pilot's decision-making process during the event. For that purpose, a comprehensive set of scenarios was developed based on an analysis of VRS-related accidents from the National Transportation Safety Board (NTSB) database. Overall, the scenarios successfully triggered VRS encounters as well as avoidance actions, providing insight into pilots' behavior in and around VRS conditions.
Sotiropoulos-Georgiopoulos, Eleni, Mavris, Dimitri, Johnson, Charles, Payan, Alexia
Lighting and Visual Information for Vulnerable Road User (VRU) Safety: An Introductory ReviewVRUSC-002-20245/3/2024
Injuries and fatalities among pedestrians, cyclists, scooterists, highway road workers, and safety and emergency personnel—often referred to as vulnerable road users (VRUs)—continue to rise at alarming rates worldwide. Emphasizing the urgent need for enhanced safety measures, this study, commissioned by the Vulnerable Road User Safety Consortium™ (VRUSC) and conducted by the Light and Health Research Center at the Icahn School of Medicine at Mount Sinai, evaluates the potential effectiveness of lighting and visual information systems in improving VRU safety. The white paper presents a synthesis of published research on lighting and markings from the perspective of both human drivers and machine vision systems. It identifies potential preliminary guidelines for the intensity, color, temporal, and spatial characteristics of lighting and visual information that can help prevent crashes involving VRUs. The study also highlights knowledge gaps that require further research, including the development of standardized lighting specifications tailored to the unique needs of VRUs.
Vulnerable Road Users Safety Consortium
Brake Pad Life Monitoring System Using Machine Learning2024-26-00321/16/2024
In the context of vehicular safety and performance, brake pads represent a critical component, ensuring controlled driving and accident prevention. These pads consist of friction materials that naturally degrade with usage, potentially leading to safety issues like delayed braking response and NVH disturbances. Unfortunately, assessing brake pad wear remains challenging for vehicle owners, as these components are typically inaccessible from the outside. Moreover, Indian OEMs have not yet integrated brake pad life estimation features. This research introduces a hybrid machine learning approach for predicting brake pad remaining useful life, comprising three modules: a weight module, utilizing mathematical formulations based on longitudinal vehicle dynamics to estimate vehicle weight necessary for calculating braking kinetic energy dissipation; and temperature and wear modules, employing deep neural networks for predictive modeling. Notably, the model’s training leverages rig-level data, with limited vehicle-level data for validation, achieving a validation accuracy of 94.8%. This innovative indirect approach holds the potential to be deployed universally in vehicles, enhancing safety without imposing additional burdens on customers or the environment.
Iqbal, Shoaib, Bhambri, Mihir, lahase, Rahul
A Fast Search Method for Edge Hazardous Scenarios Based on Semi-Supervised Anomaly Detection2023-01-705712/20/2023
Finding edge hazardous scenarios which appear very infrequently in the dataset than common hazardous scenarios is essential for implementing scenario-based testing of autonomous driving systems(ADs). However, it is difficult to evaluate the rarity of dynamic scenarios with huge scenario space high-dimensional time series, making it difficult to search for edge hazardous scenarios quickly. To solve this problem, this paper proposes a Semi-supervised anomaly detection method combining MiniRocket and DAGMM(Semi-MiniRocket-GMM, SRG), which treats edge hazardous scenarios as anomalous samples of common hazardous scenarios. SRG uses a small number of samples of common hazardous scenarios to guide interpretable feature extraction and clustering of a large amount of high-dimensional unlabeled temporal data and finds rarer edge hazardous scenarios based on anomaly evaluation to improve the coverage of test scenarios. The method is validated in the open-source natural driving dataset HighD. Compared with DAGMM, the SRG method can find edge hazardous lane change scenarios more quickly and accurately with a few samples of hazardous scenarios. The SRG method aimed at discovering edge hazardous scenarios can both guide the direction of generating scenarios and speed up the testing process.
Li, Mengyu, Li, Fang, Guo, Zihan, Wang, Lifang
PROVIZIO eyes safety with radar, not lidar23AUTP12_0212/1/2023
Provizio promises its 5D Perception stack can safely compete with expensive lidar sensors at a fraction of the cost. “Safety first” is more than a catchphrase. For sensing company Provizio, it's the only way the transportation industry should introduce autonomous vehicles. In Provizio's view, using AV building blocks - technology such as automatic emergency braking and lane-keep assist - can be valuable in ADAS systems, but they should not be used to drive vehicles until the perception problem has been solved. “It's not that we're skeptical about autonomous driving, it's just that we strongly believe that the industry has taken this wrong path,” Dane Mitrev, machine learning engineer at Provizio, told SAE Media at September 2023's AutoSens Brussels conference. “The industry has looked at things the other way around. They tried to solve autonomy first, without looking at accident prevention and simpler ADAS systems. We are building a perception technology which will first eliminate road fatalities, then deliver next-generation ADAS - and then solve autonomy. We believe that's the right order.”
Blanco, Sebastian
BRINGING ROBOTIC PLATFORMS FROM VEHICLE TESTING TO WARRIOR TRAINING2024-01-40578/10/2023
ABSTRACT In order to expedite the development of robotic target carriers which can be used to enhance military training, the modification of technology developed for passenger vehicle Automated Driver Assist Systems (ADAS) can be performed. This field uses robotic platforms to carry targets into the path of a moving vehicle for testing ADAS systems. Platforms which are built on the basis of customization can be modified to be resistant to small arms fire while carrying a mixture of hostile and friendly pseudo-soldiers during area-clearing and coordinated attack simulations. By starting with the technology already developed to perform path following and target carrying operations, the military can further develop training programs and equipment with a small amount of time and investment. Citation: M. Bartholomew, D. Andreatta, P. Muthaiah, N. Helber, G. Heydinger, S. Zagorski, “Bringing Robotic Platforms from Vehicle Testing to Warrior Training,” In Proceedings of the Ground Vehicle Systems Engineering and Technology Symposium (GVSETS), NDIA, Novi, MI, Aug. 15-17, 2023.
Bartholomew, Meredith, Andreatta, Dale, Muthaiah, Ponaravind, Helber, Nick, Heydinger, Gary, Zagorski, Scott
Adaptive Headlamp and Side Mirror Using Linear Regression Based on Raspberry Pi3SAE-PP-003507/31/2023
The greatest number of fatal traffic accidents occurs on curved roads at nighttime. In most cases, the late recognition of objects in the traffic zone plays an important role. This highlights to the importance of the role of automobile forward lighting systems. This paper developed a proto-type auto adjustable headlamp and mirror tilt to improve cost and reliability. Also, an adaptive mirror is implemented to remove the blind spots while taking turns. The methodology used here adaptive headlamps and mirrors are developed using Raspberry Pi3 as hardware and Python is used as programming language. Machine learning algorithm “Linear regression” is used for computing the output. Machine Learning Linear regression is considered here as it simple and efficient algorithm in terms of implementation and memory usage. Easily available components like Raspberry Pi3, LDR Sensor, ADXL Gyroscope are used and the design is developed to provide the steering mechanism for the headlamps and mirror which are actuated along with the steering of the front wheels. Around 15% increase in the illuminated area on road and 20% increase in the side mirror view is achieved.
Ekatpure, Rahul
When and How to Apply Automatic Emergency Brakes Based on Risk Perception and Professional Driver Emergency Braking Behavior10-07-04-00287/26/2023
The key issues of automatic emergency braking (AEB) control algorithm are when and how to brake. This article proposes an AEB control algorithm that integrates risk perception (RP) and emergency braking characteristics of professional drivers for rear-end collision avoidance. Using the formulated RP by time to collision (TTC) and time headway (THW), the brake trigger time can be determined. Based on the professional driver fitting (PDF) characteristic, the brake pattern can be developed. Through MATLAB/Simulink simulation platform, the European New Car Assessment Programme (Euro-NCAP) test scenarios are used to verify the proposed control algorithm. The simulation results show that compared with the TTC control algorithm, PDF control algorithm, and the integrated PDF and TTC control algorithm, the proposed integrated PDF and RP control algorithm has the best performance, which can not only ensure safety and brake comfort, but also improve the road resource utilization rate.
Lai, Fei, Huang, Chaoqun, Jiang, Chengyue
A Multiple-Input Multiple-Output Radar-Based Rider Assistance System for Personal Light Electric Vehicles2023-01-50407/6/2023
The use of personal light electric vehicles (PLEVs), such as electric scooters, has rapidly increased in recent years. However, their widespread use has raised concerns about rider safety due to their vulnerability in shared traffic spaces. To address this issue, this paper presents a radar-based rider assistance system aimed at enhancing the safety of PLEV riders. The system consists of an adaptive feedback system and a single-channel anti-lock braking system (ABS). The adaptive feedback system uses multiple-input multiple-output (MIMO) radar sensors to detect nearby objects and provide real-time warnings to the rider through haptic, visual, and acoustic signals. The system takes into account traffic density and uses online data to warn about obscured objects, thereby improving the rider’s situational awareness. Results from testing the feedback system show that it effectively detects potential collisions and provides warning signals, reducing the risk of accidents. The ABS is designed to prevent dangerous braking scenarios in single-track vehicles, such as rear-wheel lift-off and front-wheel locking. A virtual model was created to simulate critical riding situations and determine suitable control parameters. Testing of the MiniMAB ABS in real road tests using these parameters showed that it effectively prevented rear-wheel lift-off on high-grip roads and front-wheel locking on low-friction surfaces during emergency braking, improving riding stability and steerability. In conclusion, the results of this study indicate that the use of the proposed rider assistance system has the potential to greatly contribute to the safe and conflict-free shared use of traffic spaces. The system provides real-time warnings to the rider, thereby reducing the risk of accidents. The implementation of the ABS improves riding stability and steerability, providing a safer and more pleasant riding experience. The system offers a new and improved solution to the growing concerns surrounding the safety of PLEV riders.
Pyschny, Jan, Berger, Felix, Rothen, Samuel, Denker, Joachim, Frantzen, Michael, Roder, Felix, Kneiphof, Simon
Automatic Emergency Collision Avoidance of Straight-Crossing Intelligent Vehicle at a 90-Degree Intersection Based on Vehicle-to-Everything Technology09-11-03-00104/27/2023
In order to reduce collision at a 90-degree intersection, an automatic emergency collision avoidance control method for intelligent vehicles based on vehicle-to-everything (V2X) technology is proposed. Most of the existing automatic emergency braking (AEB) control algorithms are designed for a single high-friction road with reference to the European New Car Assessment Programme (Euro NCAP) evaluation procedures, and they do not consider changes in road friction. Thus, it may be difficult to avoid collision successfully on a low-friction road. Although some studies have considered the variation of road friction, they are only applicable to straight-line rear-end collisions and cannot be directly applied to intersections. In addition, most studies regard the vehicle only as a particle, ignoring the actual dynamic characteristics of the vehicle. The main contribution of this article is to present an AEB control strategy by V2X technology, which can make the intelligent vehicle avoid collisions at a 90-degree intersection effectively. The proposed time-to-collision (TTC) adaptive algorithm has considered various road surfaces, and its effectiveness is verified by the co-simulation of Matlab/Simulink, CarSim, and Prescan on a typical urban intersection road.
Lai, Fei, Yang, Hui, Huang, Chaoqun
Active Safety Roadside Concrete Divider Surrogate RecommendationJ3234/2_202303 (Current)3/8/2023
The scope of this document is to provide the design specifications/requirements/guidelines for concrete divider surrogates that represent actual concrete dividers to the in-vehicle sensors and can be used for performance assessment of such in-vehicle sensing systems in real-world test scenarios/conditions. Therefore, this document only includes the recommended concrete divider surrogate characteristics for automotive cameras, LiDARs, and/or radars. Concrete dividers are also known as concrete barriers [1].
Active Safety and Driver Support Systems Standards Committee
Volvo's electric high-wire act23AUTP03_143/1/2023
Volvo calls its all-new EX90 SUV the safest and most technically adept model in the company's 95-year history, which includes such achievements as the world's first three-point automotive seat belt in 1959. Even before this luxury EV logs its first mile on global roads that take more than 1 million human lives every year, Volvo asserts the EX90 will eliminate up to one in five serious injury accidents, and one in 10 accidents overall. That claim is based not on fuzzy math, said Lotta Jakobsson, a 33-year company veteran and specialist in injury protection, but on Volvo's industry-unique accident database that's been a wellspring of company safety innovations since the 1970s.
Ulrich, Lawrence
Model-Based Systems Engineering of the Aft Collision Assist Advanced Driver Assistance System15-16-03-00122/13/2023
The Aft Collision Assist (ACA) is an Advanced Driver Assistance System (ADAS) that is added to a vehicle and integrates with the native systems of that vehicle. The ACA is used to monitor and reengage a distracted driver of an approaching vehicle that the ACA system calculates will imminently rear-end the host vehicle. This work provides a brief overview of existing ADAS that perform similar functions, the regulatory statutes and requirements that impact the ACA functionality, and Model-Based System Engineering (MBSE) model diagrams of the ACA. The MBSE model diagrams presented are State Machine, Conceptual Data Model, Use Case, System Requirements, and Regulatory Requirements for the entire ACA system. The MBSE models and regulatory constraints presented within are used to refine and specify the ACA method of attracting a distracted driver’s attention.
Rictor, Andrew, Chandrasekar, Chandra V.
A Concise Camera-Radar Fusion Framework for Object Detection and Data Association2022-01-709712/22/2022
Multi-sensor fusion strategies have gradually become a consensus in autonomous driving research. Among them, radar-camera fusion has attracted wide attention for its improvement on the dimension and accuracy of perception at a lower cost, however, the processing and association of radar and camera data has become an obstacle to related research. Our approach is to build a concise framework for camera and radar detection and data association: for visual object detection, the state-of-the-art YOLOv5 algorithm is further improved and works as the image detector, and before the fusion process, the raw radar reflection data is projected onto image plane and hierarchically clustered, then the projected radar echoes and image detection results are matched based on the Hungarian algorithm. Thus, the category of objects and their corresponding distance and speed information can be obtained, providing reliable input for subsequent object tracking task. Results shows that the fusion method greatly improves the perception dimension and accuracy of intelligent vehicles in adverse environments, its matching accuracy reaches 62.3% on the VTTI dataset and the camera-radar association process takes 0.013s per frame. All implementation is based on ROS (Robot Operation System) to facilitate the feasible application of algorithms.
He, Yingjie, Zhao, Jian, Lyu, Nana, Li, Linhui, Liu, Pengbo
Perception-Aware Path Planning for Autonomous Vehicles in Uncertain Environment2022-01-707712/22/2022
Recent researches in autonomous driving mainly consider the uncertainty in perception and prediction modules for safety enhancement. However, obstacles which block the field-of-view (FOV) of sensors could generate blind areas and leaves environmental uncertainty a remaining challenge for autonomous vehicles. Current solutions mainly rely on passive obstacles avoidance in path planning instead of active perception to deal with unexplored high-risky areas. In view of the problem, this paper introduces the concept of information entropy, which quantifies uncertain information in the blind area, into the motion planning module of autonomous vehicles. Based on model predictive control (MPC) scheme, the proposed algorithm can plan collision-free trajectories while actively explore unknown areas to minimize environmental uncertainty. Simulation results under various challenging scenarios demonstrate the improvement in safety and comfort with the proposed perception-aware planning scheme.
Chen, Zhan, Xiong, Lu, Tang, Chen
Items per page:
1 – 50 of 133