Browse Topic: Driver assistance systems

Items (850)
This paper investigates the integration of Artificial Intelligence (AI) within radar-based perception for Advanced Driver Assistance Systems (ADAS) under safety considerations aligned with ISO 26262 [1] for functional safety and ISO 21448 (SOTIF) [2] for performance-related safety of the intended functionality. The study evaluates a hybrid architecture in which AI-based perception modules are combined with deterministic supervisory mechanisms to maintain safety compliance. A simulation-based case study using CARLA with radar sensor modeling is presented to compare a deterministic radar perception pipeline with an AI-enhanced approach under nominal and degraded environmental conditions. Performance is evaluated using precision, recall, and F1 score metrics. Results indicate improved recall and F1 score under adverse scenarios for the AI-based perception module, accompanied by a moderate increase in false positives. The paper discusses architectural constraints required to limit non-deterministic behavior, including confidence gating, deterministic supervision, and scenario-based validation. The findings are limited to simulation and are intended to provide preliminary insights into the technical and safety implications of incorporating AI-based radar perception within ISO 26262-compliant ADAS architectures.
Jain, Yesha
Ultrasonic sensors are widely deployed in automotive driver assistance systems for near-range environment perception and provide safety-relevant inputs for functions such as parking assistance and automated parking. With increasing vehicle automation, the integrity and availability of ultrasonic sensor data become more critical, as compromised measurements may lead to incorrect vehicle decisions and hazardous behavior. While prior research has extensively studied physical attacks on ultrasonic sensors, a structured cybersecurity risk analysis in accordance with automotive cybersecurity standards, combined with experimental validation, is largely missing. In particular, the communication interface between ultrasonic sensors and control units has received limited attention despite its relevance as a potential attack surface. This paper presents a systematic security analysis of an automotive ultrasonic sensing system based on a demonstrator setup. The work applies a Threat Analysis and Risk Assessment methodology aligned with ISO/SAE 21434 and HEAVENS 2.0 to identify security-relevant assets, threat scenarios, and attack paths. Risk levels are derived by evaluating potential impact and attack feasibility. To validate the risk assessment, a structured test strategy is developed using the ISTQB test process and translated into laboratory experiments. Both digital attacks targeting the sensor communication interface, with DSI3 selected as the representative protocol, and physical manipulations of the sensor environment are examined. Experimental results show that selected communication-level attacks can be realized with moderate effort and can cause controlled falsification or loss of measurement data. Physical environmental manipulations significantly degrade signal quality but do not fully suppress object detection in the evaluated configuration. The findings largely confirm the initial risk assessment while enabling refinement of attack feasibility parameters. The results provide a validated linkage between automotive cyber-security risk assessment methods and practical testing of ultrasonic sensing systems and underline the importance of jointly addressing communication interfaces and physical effects in future security concept development.
Gahm, SebastianHaller, JonathanKriesten, Reiner
The development and validation of advanced driver-assistance systems (ADAS) and automated driving systems (ADS) are shifting from traditional linear V-model processes toward more iterative engineering cycles. Despite faster iteration, these safety-critical systems remain subject to stringent regulations. Standards and guidance, including UNECE UN Regulation No. 157 and ISO/TS 5083, emphasize traceability, transparency, and explainability throughout development and validation. Nevertheless, as ADAS/ADS are developed and validated in faster, more iterative release cycles, additional stakeholders become involved and new explainability requirements emerge. These requirements vary between stakeholders and across development, validation, and post-market deployment phases, yet they are not systematically captured in the current state of research and practice. Therefore, to ensure that explainability supports rapid iteration, it is essential to identify relevant stakeholders and specify their explainability needs. Standards such as IEEE Standard 7001-2021 provide a broad foundation for transparency in autonomous systems. However, their generic nature does not address the domain-specific complexities of ADAS/ADS. Furthermore, a conceptual gap remains between general transparency principles and explainability requirements in automotive development and validation. Building on IEEE Standard 7001-2021, this paper first offers a stakeholder taxonomy in the context of ADAS/ADS, then proposes a stakeholder-oriented analysis of explainability requirements within an automated driving use case and contexts. This analysis specifically focuses on the motivations for requiring explainability and the necessary explanation modalities. Finally, the paper discusses the limitations of the analysis and outlines directions for future research. The results of the paper provide a structured guideline for stakeholder-oriented explainability requirements in ADAS/ADS.
Liu, XuanhengBairy, AkhilaPaudel, BijayAdolph, LaurenzHeck, MelanieHettich, LennardNägele, Ann-ThereseRudolf, KorbinianBause, KatharinaDüser, TobiasSchwammberger, Maike
Automotive Engineering: June 202626AUTP066/4/2026
New York 2026: diversity on full display New powertrain choices keep popping up on new vehicles from OEMs that debuted at NYIAS this year. Sealing integrity in a Formula 1 limited-slip differential High-temperature hydraulic control in a Formula 1 drivetrain requires dimensional stability, controlled sealing force, and resistance to wear under sustained pressure cycling. Inside the limited-slip differential, the sealing architecture plays a defined mechanical role in maintaining consistent torque management under race conditions. From ADAS to autonomy How engineering thermoplastics can advance sensor-based technologies. Synthetic data and the future of ADAS validation Why ADAS validation can't be solved with more miles alone. Intelligent power distribution will change the way vehicles are designed Electronic fuse (eFuse) technology can create electronic power distribution modules (ePDMs) for architectural flexibility, higher reliability, greater safety, and proactive maintenance. Editorial Maybe more than ever, let's talk transportation diversity The Navigator Can legacy automakers finally succeed with SDVs? AI scares and excites cybersecurity professionals at WCX Expert claims war hurting China's already-struggling economy NHTSA open to negotiated rulemaking on some safety issues Resilient propulsion strategies require options Driven: Honda Fastport eQuad Prototype Product Briefs Spotlight: Connectors & harnesses, EV thermal management Q&A Neural Concept's Thomas von Tschammer: Working with AI at speed
In the two months since Microvision bought Luminar and acquired key tech and talent, the sensor company has been busy. In that time, they've merged key lidar units from each company and created a perception software stack to run it in a convincing demo of its ADAS and autonomous capabilities. The company is also pushing innovative lidar tech into the defense drone and antidrone markets, already working with a German defense supplier that works with NATO member countries.
Clonts, Chris
Letter from the Guest Editor
Tylko, Suzanne
Roadway departures remain a major cause of crashes, injuries, and fatalities on U.S. roads. Technologies such as lane keeping assist (LKA) and lane centering assist (LCA) can help mitigate these crashes, but their development involves extensive characterization of the parameter space in which they operate. Lane and road departures (LDs/RDs) and lane changes (LCs) must be systematically described and quantified to distinguish kinematic features, identify contributing factors, and benchmark system influence on lateral control. This study developed a unified pipeline to mine over 36 million miles of naturalistic driving study (NDS) data collected from more than 3800 participants. The pipeline integrates various types of signals to detect roadway boundary crossings, classify LKA-relevant scenarios, and extract roadway, driver, environmental, and assistance-related parameters. Lane keeping epochs with and without LKA were also extracted to quantify system influence on lateral control. In the NDS analysis, crashes include both object contact events and RDs, defined as non-premeditated departures from the intended travel surface involving at least one tire. Analysis of pre-identified crashes in the NDS showed that unintentional RDs accounted for 5.67%, unintentional LDs for 1.76%, and intentional LCs for 1.55%, corresponding to lower-bound rates of 2.7, 0.8, and 0.7 crashes per million vehicle miles traveled. RD crashes were predominantly right-sided, LD crashes left-sided, and both were overrepresented on curves and under adverse conditions. Loss of control preceded 22% of RD crashes and 69% of LD crashes. Beyond crashes and near-crashes (CNCs), the algorithm identified approximately 3 million LCs and 0.3 million LDs/RDs. LCs typically involved larger crossing angles that decreased with speed, while departures clustered within 0°–2°. Compared with CNCs, these occurred at higher speeds and smaller angles. LKA consistently reduced lateral variability without biasing the mean offset.
Ali, GibranTerranova, PaoloWilliams, VickiHolley, DustinSaffy, JoshuaAntona-Makoshi, JacoboKefauver, KevinShull, EmilyLi, EricVenegas, Michael
This research examined the performance of SAE Level 2 (L2) advanced driver assistance systems (ADAS) in crash-imminent scenarios (CIS), with particular attention to how vehicle configuration like body style and powertrain (internal combustion engine, plug-in hybrid, electric vehicle) influences vehicle system performance. The objectives were to (1) identify CIS relevant to L2-equipped vehicles using crash databases and naturalistic driving studies (NDSs), (2) develop scenario-based test procedures and test matrices, and (3) evaluate system and vehicle responses across configurations and conditions. Multiple crash data sources were analyzed, including NHTSA’s Standing General Order dataset of L2-related crashes, the Fatality Analysis Reporting System, the Crash Report Sampling System, and NDS data from the Second Strategic Highway Research Program and the Virginia Tech Transportation Institute L2 NDS. Coded variable analyses from the datasets identified three common CIS: lane and road departures, rear-end striking events, and intersection conflicts. Supporting variables such as speed, roadway condition, and driver actions were also extracted to characterize scenarios and inform test development. Tests were executed at a closed-track testing facility using four vehicles selected for diversity of L2 systems, body types, and powertrains. Phase 0 exploratory testing assessed vehicle kinematics and L2 responses to refine the test matrix. Phases 1 and 2 conducted controlled evaluations of selected CIS, with expansion factors reflecting real-world crash variability. The testing highlighted interactions between L2 features and active safety systems. For example, results showed that all four vehicles employed distinct hand-off strategies between L2 longitudinal control and active safety systems during rear-end striking crash scenarios, and AEB engagement was strongly correlated with TTC at the moment the vehicle identified the crash partner. This work contributes novel insights into vehicle L2 and ADAS behavior in CIS events across multiple factors and provides a structured framework to evaluate system behavior for those crash-imminent scenarios.
Beale, GregoryKefauver, KevinVenegas, MichaelLi, EricChen, JayHuggins, StevenGuduri, BalachandarLlaneras, Eddy
Drivers frequently encounter Type II dilemma zones at signalized intersections, where the decision to stop or proceed during the onset of a yellow indication can be ambiguous. Decision-making relies on drivers’ expectations of the yellow change interval duration and behavioral factors. While boundaries of these zones are well studied, less is known about how familiar drivers are with their local yellow indication laws, which vary from state to state, and whether their typical reactions to yellow indications align with the laws. Existing interventions like signal timing adjustments, improved vehicle detection, and advance warning signs reduce the number of drivers caught in dilemma zones but may not reach distracted drivers. In-vehicle alerts tailored to dilemma zone scenarios are a potential solution not yet implemented widely in North America. This study addresses how drivers may interpret these alerts. A web-based survey of 640 licensed drivers in Michigan and Washington (ages 18–85) assessed respondents’ knowledge of their state’s law, typical responses to yellow indications, interpretations of proposed in-vehicle alerts, and preferences for alert modality, frequency, and placement. These states were selected for their differing yellow indication laws—restrictive in Michigan, permissive in Washington. Nine alert visuals were tested, including pairs of implicit and explicit messages, and were inspired by or designed to address gaps in prior research. Respondents evaluated these alerts in response to hypothetical intersection scenarios that varied by the presence of other vehicles. Results revealed a prevalent misunderstanding of local yellow indication laws across both states. Statistical analyses showed significant differences in rankings among the nine alert visuals, and explicit messages showed higher rates of correct interpretation. Findings show overall driver support for dilemma zone alerts, but higher receptivity in drivers who more frequently use other ADAS features and lower receptivity in drivers within older, but not the oldest, age groups. Future research could explore whether these alerts promote safe behaviors aimed at crash avoidance.
Anderson, ErikaJashami, HishamAhmed, AnannaHurwitz, David
Vehicle maneuver data are essential for perception and planning in advanced driver-assistance systems (ADAS) and automated driving systems (ADS). While high-quality annotations improve machine-learning performance, existing maneuver datasets remain fragmented, labor-intensive to annotate, and inconsistent in semantic richness. Challenges persist in scalability, interpretability, and contextual labeling. This article establishes a structured framework for maneuver data analysis by combining a systematic review of existing resources with the development of a new multimodal dataset. First, we conduct a systematic review of publicly available datasets such as HDD, KITTI, BDD-X, D2CAV, Brain4Cars, DrivingDojo, and the Driving Behavior Database. We further evaluate the data modality and sensor configurations including event data recorders, onboard logging systems, and smartphone sensing. We then propose the Matt3r Data Collection System with modern metadata management, which integrates video, GPS, and IMU signals into temporally coherent clips. Next, we outline the limitations of traditional annotation approaches, which rely on manual labeling and rule-based methods. To address the limitations of traditional manual and semi-automated labeling, we propose a Vision–Language Model (VLM)–driven annotation pipeline. VLMs generate maneuver categories and causal explanations through prompt-based reasoning, with selected outputs refined through human-in-the-loop verification. Finally, we propose an annotation quality evaluation based on accuracy, inter-annotator agreement, credibility, consistency, and efficiency gain. In summary, this article bridges the gap between the environment perception requirements of existing ADAS and ADS systems and the developing capabilities of generative artificial intelligence. By providing a novel and scalable research approach for AI-driven maneuver data annotation and analysis, this article supports data engineering efforts for both research and practical applications aimed at enhancing vehicle safety.
Bai, LingYuan, ChongyuOsman, IslamLin, ZiruiMirab, GhazalSaheb, AmirParnian, NedaShapiro, EvgenyShehata, Mohamed S.Liu, Zheng
Building a trusted digital twin and decision-centric simulation ecosystem The automotive industry has been experiencing significant change and transformation. Electrification, software-defined vehicles, advanced driver assistance systems, and increasing electrical system integration are fundamentally reshaping how vehicles are designed and validated. As integration complexity continues to increase, the expectations for design cycle times are being compressed. Programs that once relied on extended validation timelines are now expected to deliver the same level of confidence in a fraction of the time. Traditional engineering workflows were built around sequential design phases, iterative simulations, and heavy reliance on physical validation. Design concepts were documented, prototypes were constructed, tests were performed, and results were compiled in reports and specifications that informed the next iteration. That approach worked well when systems were less complex and product life cycles were longer. In recent years, the volume of data, the speed of development, and the interconnected nature of modern vehicle architectures demand a different approach.
Patterson, Jeremy
Edge detection is fundamental for intelligent vehicle applications, directly supporting ADAS functions such as lane detection, obstacle recognition, and scene understanding. The conventional Canny edge detection method exhibits notable shortcomings, especially in color-image processing, adaptive threshold selection, and preserving edge integrity under noisy conditions. In this study, we present an enhanced Canny edge detection framework tailored for ADAS-oriented intelligent vehicle systems, incorporating a quaternion-based weighted averaging scheme for color preservation, adaptive thresholds derived from gradient-amplitude histograms, multiscale edge localization via scale multiplication, and a novel gravitational-field-intensity operator for improved gradient robustness. Moreover, we extend the method to vanishing-point estimation an essential ADAS capability by performing precise intersection calculations combined with clustering techniques such as DBSCAN and RANSAC. Experimental evaluations demonstrate that the proposed algorithm markedly outperforms traditional approaches in edge clarity, localization accuracy, and noise resilience, underscoring its promise for strengthening ADAS perception modules in intelligent vehicles.
Uppala, Rohit RajKaye, MuraliZadeh, MehrdadTan, Teik-Khoon
Achieving full vehicle autonomy is not just about adding sensors or compute - it requires a fundamental shift in how vehicles are architected. Autonomous systems rely on higher-resolution sensors, massive processing power, and the ability to fuse data from multiple sources in real time. Centralized in-vehicle architectures, which consolidate computing and enable sensor fusion, place unprecedented demands on connectivity. Precise time synchronization across systems becomes critical, as does advanced control to ensure safe and reliable operation. Any delay or data loss can impact decision-making, making robust, resilient communication links essential. High-performance connectivity is the backbone of this evolution. It must deliver the highest bandwidth to handle massive streams of sensor data, support long-reach connections across the vehicle, and maintain error-free performance even in the most challenging electromagnetic environments. This combination of speed, reach, and reliability forms the foundation that enables higher-level ADAS and ultimately autonomous driving to move from concept to reality.
Shwartzberg, Daniel
Ensuring ISO 26262 functional safety in advanced driver assistance systems (ADAS) is increasingly complex as these platforms integrate artificial intelligence (AI) for perception, decision-making, and vehicle control. Traditional safety mechanisms are largely deterministic, but AI introduces non-determinism, creating challenges for verification, validation, and certification. Real-time vehicle telemetry, sensor outputs, and environmental inputs are processed through machine learning algorithms that forecast hardware and software faults before they escalate into hazardous conditions. These predictions are systematically integrated with ISO 26262 safety measures, enabling adaptive diagnostics, fault isolation, and rapid recovery strategies. The AI model introduces hazards such as data bias, model drift, opaque decision-making, and unsafe automation. A dedicated AI Hazard Analysis and Risk Assessment addresses data quality, validation, monitoring, explainability, and fail-safe mechanisms alongside system-level safety controls. The proposed approach demonstrates measurable improvements, including up to 25 % higher diagnostic coverage and fault-recovery times under 30 ms, while maintaining ASIL-D compliance and adhering to FTTI, SPFM, and DC requirements. Hardware-in-the-loop (HIL) simulations validate system performance and robustness under diverse operational scenarios. Future work focuses on uncertainty quantification and explainable AI integration, enhancing traceability and safety certification readiness for intelligent ADAS controllers. By demonstrating how AI can complement functional safety principles instead of conflicting with them, this study provides OEMs and Tier-1 suppliers with a roadmap for deploying certifiable, intelligent, and resilient ADAS platforms. This framework ensures safer, more reliable AI-enhanced vehicle systems while bridging the gap between emerging AI technologies and rigorous functional safety standards. This paper presents a predictive fault management framework that enhances functional safety in ADAS controllers by combining AI-driven predictive models with ISO 26262 safety mechanisms. This work uniquely bridges deterministic ISO 26262 workflows with predictive AI fault forecasting. In this framework, the AI model is used solely as a diagnostic enhancement and is not credited as an ISO 26262 safety mechanism; all safety decisions and fault reactions remain under deterministic safety-shell control.
Abdul Karim, Abdul Salam
Embedded vision systems are essential for contemporary applications, including robotics, advanced driver assistance systems (ADAS), and intelligent surveillance; yet they frequently experience diminished image quality due to resource constraints, environmental variability, and inconsistent illumination conditions. Such degradations impact multiple visual attributes—sharpness, contrast, color accuracy, noise levels, and structural similarity—that are critical for reliable perception in safety- and performance-driven domains. This study introduces a comprehensive system-level calibration architecture that integrates three coordinated layers: sensor-level adjustment, firmware optimization, and adaptive software enhancements. At the sensor level, exposure control, gain tuning, and white balance adjustments mitigate luminance imbalance and color shifts under changing light conditions. Firmware optimization leverages image signal processor (ISP) parameters to reduce temporal and spatial noise, refine tone mapping, and correct color reproduction through calibrated color correction matrices. Software-level improvements apply adaptive sharpening, contrast enhancement, and gamma correction to maintain visual fidelity across diverse scenes. The proposed pipeline was evaluated on three representative embedded platforms—NVIDIA Jetson Nano, Raspberry Pi 4B, and STM32F7 MCU—covering a range of computational capabilities and power budgets. Experimental results demonstrate substantial improvements in image quality: Peak Signal-to-Noise Ratio (PSNR) increased from 24.2 dB to 31.6 dB in indoor low-light conditions, Structural Similarity Index (SSIM) improved from 0.73 to 0.88 in dynamic scenarios, and color accuracy (ΔE) was reduced to 3.1 in bright outdoor conditions. The complete calibration pipeline sustained real-time responsiveness (< 40 ms/frame) with acceptable power consumption (maximum 172 mW) and memory utilization (peak 35.7 MB). These results validate the modularity, efficiency, and robustness of the proposed method, making it well-suited for deployment in practical embedded vision applications where image quality, latency, and resource constraints must be balanced.
Indrakanti, Rama Kiran KumarVishnoi, NitinKamadi, Venkata
The increased integration of radar and vision sensors in modern vehicles has significantly improved environmental perception, safety, and automation. Nevertheless, conventional camera modules capture images in fixed, continuous frames, leading to unnecessary data processing, power consumption, and heat generation in the limited space of small sensors. The paper discusses the technology of Radar Based Dynamic Pixel Activation (RDPA); whereby radar data can be used to dynamically activate specific pixels on the camera sensor, optimizing image capture and processing. Through a systematic literature review of peer-reviewed articles published between 2021 and 2025, we examined the literature on radar-camera fusion, adaptive imaging, and sensor design that is efficient in power consumption. The review indicates a research gap that there is no current paradigm that dynamically activates sensor pixels at the hardware level using radar data. We aggregated ten topical studies and proposed a conceptual model where radar-determined Regions of Interest (ROIs) trigger localized pixel activation. The framework reduces the computational load, improves power efficiency and enhances thermal performance without affecting image fidelity. The paper also explains how RDPA may affect the Driver Monitoring Systems (DMS), Occupant Monitoring Systems (OMS), and Advanced Driver Assistance Systems (ADAS), and how it is more beneficial than traditional full frame imaging. Difficulties with synchronization, hardware interpenetration, and algorithmic synchronization are discussed. Altogether, RDPA is an excellent prospect to intelligent, energy-saving, and thermally stable vehicle perception systems of the next generation.
Kasarla, Nagender Reddy
This study presents the development and validation of a muddy water spray apparatus designed to simulate dust contamination on vehicle sensors for sensor cleaning system testing. It is important to have a constant and quantifiable test environment for the vehicle development process. For verifying the apparatus, muddy water, prepared by mixing standardized dust powder, salt, and water to maintain constant contamination test conditions, was sprayed onto glass specimens to evaluate equipment consistency. Deposited dust weight and thickness were measured across multiple spray cycles, with statistical analyses confirming consistent and reliable deposition. Paired t-tests indicated no significant difference between sample positions, demonstrating uniform spray distribution. The apparatus was further applied to individual infrared (IR) cameras to observe performance degradation under dry and wet contamination conditions showing statistically consistent increases in contamination levels. Application of the system in low-temperature performance testing of sensor cleaning systems on a development vehicle's rearview mirror yielded significant reductions in test time and variability compared to manual methods. These results affirm the apparatus as an effective, quantitative, and reproducible method for contamination simulation, contributing to streamlined and standardized sensor cleaning system development in the automotive industry.
Jinhyeok, Gong
Adaptive Cruise Control (ACC) has become a widely adopted driver-assist technology, designed primarily to regulate a vehicle’s longitudinal movement while maintaining a safe following distance from the preceding vehicle. A key performance criterion is the system’s ability to detect and respond to both moving and stationary target vehicles within the ego vehicle’s path. While manufacturers typically validate ACC performance within specific speed ranges, responding to stationary objects remains particularly challenging due to limited sensor range, difficulty in detecting distant stationary targets, and constrained deceleration capabilities. Beyond certified operating limits, overall system reliability may degrade. Nonetheless, increasing industry and regulatory expectations are driving the need to extend ACC functionality across wider and more clearly defined speed domains. Modern ACC systems are further evolving to recognize and respond to various road features, including traffic lights, STOP signs, intersections, curved road segments, and roundabouts—an expanding set of scenarios enabled by multi-sensor fusion and map integration using standard definition (SD) and high definition (HD) maps. Regulatory frameworks are increasingly addressing these map-based functionalities. This paper investigates the interaction between map-based functionalities and traditional ACC behavior, specifically examining how map integration enhances ACC responsiveness to critical scenarios. Due to the wide variety of possible cases, this study focuses on stationary vehicle encounters, recognized as the most challenging and safety-critical scenario, particularly at higher speeds. Simulation studies are conducted to evaluate the impact of map-based augmentation on ACC performance, with results demonstrating performance improvements. For instance, at 50 mph on straight roads, the ego vehicle safely stopped ~4 meters from the target stationary vehicle using map-based anticipatory braking, compared to less than 1 meter with traditional ACC. These findings highlight the extended operational capability and safety benefits offered by the proposed approach, even beyond conventional speed limits.
Awathe, ArpitPatel, DarshMathur, DhruvRaut, Abhinandan Vijay
Trust calibration is vital for safe human–automation interaction but remains largely qualitative. This study develops multiple quantitative frameworks modeling trust as a function of automation reliability. Four progressive models of binary, linear, triangular, and logistic formalize the calibrated trust zone, defining where human reliance aligns with system performance. The framework corrects major misconceptions: that trust is purely qualitative, that low trust–low reliability states are acceptable, and that overtrust and distrust pose equal risk. It establishes a minimum reliability threshold for meaningful trust and identifies distrust as the safer default in high-risk contexts. A case study on an empirical observation of 32 AI applications plotted in the trust–reliability space confirms the analysis, revealing a consistent distrust tendency where reliability exceeds user confidence and other observations. By quantifying trust through reliability, the study reframes it as a controllable safety variable, enabling predictive calibration and adaptive, trust-aware safety architectures for reliable human–AI collaboration.
Wen, HeMounir, Adil
Energy efficiency and range optimization remain critical challenges to the widespread adoption of battery electric vehicles (BEVs). As a result, there is a growing demand for intelligent driver assistance systems that can extend the operating range and reduce range anxiety. This paper presents an adaptive eco-feedback and driver rating system based on proximal policy optimization (PPO) reinforcement learning, designed to support drivers with the target to reduce energy consumption and maximize driving range. The system processes real-time driving data, such as velocity, acceleration and powertrain status. Map data of high quality is used to anticipate traffic events, including but not limited to speed limits, curves, gradients, preceding vehicles and traffic lights. This contextual awareness allows the system to continuously assess driving behavior and provide personalized, context-aware visual feedback alongside a dynamic driving behavior rating. A PPO agent learns optimal feedback strategies through continuous interaction and evaluates the impact of specific guidance actions, such as but not limited to “release accelerator pedal”, “brake” and “recuperate”, on immediate energy efficiency and long-term driver adaptation patterns. Feedback intensity and modality are dynamically tailored to individual driver profiles based on observed reaction patterns and feedback adherence. This approach encourages drivers to prioritize energy efficiency while aiming to minimize cognitive distraction and discomfort. The algorithm is implemented and validated within a driving simulation environment that replicates diverse and realistic conditions. Virtual driving tests conducted in various scenarios, such as congested urban areas, suburban routes, mountain roads and highways demonstrate that the proposed PPO-based eco-driving assistance system can reduce energy losses by about 28% compared to conventional driving behavior.
Stocker, ChristophHirz, MarioMartin, MichaelKreis, AlexanderStadler, Severin
This article investigates the optimization problem of fuel economy for heavy-duty commercial vehicles. A Dynamic Programming–Based Fuel-Saving Predictive Cruise Control (DP-FSPCC) method is proposed, which is based on the Bellman optimality principle and uses the cost function to evaluate the optimal feedback control gain, thereby improving the fuel economy of heavy-duty commercial vehicles on complex roads with varying slopes. To address the issues of low accuracy in road feature representation and poor adaptability to different driving conditions in existing slope reconstruction algorithms, the road ahead is dynamically segmented for high-precision processing by integrating ADASIS (Advanced Driver Assistance Systems Interface Specifications) map information with significant turning point detection and dynamic sensitivity analysis. An engine fuel consumption mapping model based on local gradient information is established to provide an accurate cost function for dynamic programming. Furthermore, a feedforward optimization mechanism based on slope classification is proposed. This mechanism adopts a differentiated cost function weight design strategy for different road conditions, making the control strategy more in line with actual driving experience, effectively reducing the computational complexity of dynamic programming and improving the real-time performance and optimization efficiency of the algorithm. Finally, through numerical simulations and real-vehicle tests on highways, the effectiveness and superiority of the proposed method are verified.
Jin, DapengShuai, YueWu, XinJia, TongQiao, ZhiyuanChang, ShiweiMu, Tong
Microchip Technology and Hyundai Motor Group recently announced a collaboration to test 10BASE-T1S Single Pair Ethernet (SPE) technology for advanced in-vehicle networks to provide improved ADAS and connected-vehicle features. HMG told SAE Media it is working with multiple technology partners to review the overall applicability of 10BASE-T1S technology and hopes 10BASE-T1S can help optimize the deployment of gateways and switches. The technology's ethernet-based networking concepts might also contribute to simplifying network design and implementation for future zonal architectures. We also spoke with Matthias Kaestner, corporate vice president of Microchip Technology's data center, networking and automotive business units, about the partnership, via email.
Blanco, Sebastian
Dooring accidents occur when a vehicle door is opened into the path of an approaching cyclist, motorcyclist, or other road user, often causing serious collisions and injuries. These incidents are a major road safety concern, particularly in densely populated urban areas where heavy traffic, narrow roads, and inattentive behavior increase the likelihood of such events. To address this challenge, this project presents an intelligent computer vision based warning system designed to detect approaching vehicles and alert occupants before they open a door. The system can operate using either the existing rear parking camera in a vehicle or a USB webcam in vehicles without such a feature. The captured live video stream is processed by a Raspberry Pi 4 microprocessor, chosen for its compact size, low power consumption, and ability to support machine learning frameworks. The video feed is analyzed in real time using MobileNetSSD, a lightweight deep learning object detection model optimized through TensorFlow Lite to ensure smooth and efficient processing even on resource- constrained hardware. Detected objects are classified, and the relative distance of approaching vehicles is estimated based on bounding box dimensions and simple geometric calculations. If a vehicle is detected within a predefined safety distance, the system immediately displays a clear visual warning on an in-vehicle screen, giving occupants enough time to delay opening the door and avoid a potential collision. The system was successfully implemented and tested on both a laptop and Raspberry Pi, demonstrating high accuracy, low latency, and minimal hardware requirements, making it cost effective and scalable. Looking forward, the design allows for future upgrades such as automatic door locking when a hazard is detected, audio and haptic alerts for greater situational awareness, integration with other vehicle sensors for improved detection accuracy, and seamless incorporation into commercial advanced driver assistance systems, providing a practical, affordable, and effective solution to enhance road safety and protect vulnerable road users.
C, JegadheesanT, KarthiGurusamy, Varun SankarBalraj, TharunMurugaiya, Tamilselvan
Treat foundational AV safety like seatbelts - make it non-proprietary and universal. An open safety stack, shared scenarios, benchmarks, and core validation tools can speed certification, reduce duplicated V&V and build public trust while preserving vendor differentiation. The bottleneck isn't compute - it's verification. Autonomous features are shipping in more vehicles and markets, but the gating factor is no longer raw compute. It's whether developers and regulators can verify systems against requirements and validate them against real-world operating design domains (ODDs) with confidence and repeatability. Today, many safety-critical components, from scenario libraries to pass/fail criteria, live in proprietary silos. That fragmentation slows regression testing, complicates regulator audits across regions, and duplicates effort across the industry. The result is an expensive, bespoke path to certification for every program and geography.
Musa, MohammadKhawaja, Muhammad Zain
SAE International’s Dictionary of ADAS and Connected VehiclesR-5591/20/2026
The convergence of Advanced Driver Assistance Systems (ADAS) and connected vehicle technologies is ushering in a transformative era of automotive innovation—one that is fundamentally enhancing vehicle safety, efficiency, reliability, and the overall driving experience. SAE International’s Dictionary of ADAS and Connected Vehicles stands as the definitive reference for this rapidly evolving domain, meticulously compiled to clarify and standardize the language shaping modern mobility. Inside, readers will find clear, authoritative definitions encompassing the full spectrum of technologies that enable connected and automated driving—including vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I), and vehicle-to-everything (V2X) communication systems. This comprehensive resource translates complex engineering terminology into accessible language, enabling engineers, researchers, policymakers, educators, and enthusiasts to share a common technical foundation. Reflecting the latest global standards, research, and innovations, this dictionary bridges the gap between theory and application. It fosters interdisciplinary collaboration, supports safer system design, and provides clarity for those engaged in regulatory development or technology deployment. As the pace of mobility innovation accelerates, precise and accessible communication becomes indispensable. SAE International remains committed to advancing global transportation knowledge—empowering professionals to navigate, contribute to, and shape the future of intelligent, connected, and sustainable mobility with confidence and clarity.
Quigley, Jon M.Gulve, AmolKrishnamoorthy, Jayalekshmi
Ensuring the safety and functionality of sophisticated vehicle technologies has grown more difficult as the automotive industry quickly shifts to intelligent, electric, and connected mobility. Software-defined architectures, electric powertrains, and advanced driver assistance systems (ADAS) all require strong quality assurance (QA) frameworks that can handle the multi domain nature of contemporary vehicle platforms. In order to thoroughly assess the functionality and dependability of next generation automotive systems, this paper proposes an integrated QA methodology that blends conventional testing procedures with model-based validation, digital twin environments, and real-time system monitoring. The suggested framework, which includes hardware-in-the-loop (HIL), software-in-the-loop (SIL), and over-the-air (OTA) testing techniques, concentrates on end-to-end traceability from specifications to validation. Simulating intricate situations for ADAS, electric vehicle battery temperature management, and dynamic system updates in connected platforms are prioritized. This study also outlines the main obstacles to integrating QA methods with changing regulatory environments and draws attention to discrepancies between operational performance in real-world scenarios and compliance benchmarks. Early fault detection, lifecycle validation, and continuous improvement are made possible by the QA process's transition from reactive to proactive through the integration of digital twins and predictive analytics. A strategic roadmap for QA specialists and test engineers to adjust to changing industry demands is presented in the paper's conclusion. In addition to promoting safety and dependability, the suggested framework speeds up time to market, lowers development costs, and increases consumer confidence in cutting-edge automotive technologies.
Komanduri, Arun SrinivasSrivastava, Anuj
The paper aimed to improve the accurate quantification of driver drowsiness and to provide comprehensive, evidence-based validation for a Vision-Based Driver Drowsiness and Alertness Warning System. Advanced quantification of driver drowsiness is designed to enhance distinction of true positive events from False Positive and False Negative events. Methodology to pursue this included assessing inputs such as facial features, driver visibility, dynamic driving tasks, driving patterns, driving course time and vehicle speed. The system is programmed to actively learn Eye Aspect Ratio (EAR) reference and adapt personalised EAR threshold value to process EAR frames against the learnt threshold value. This method optimized the data frames to enhance the evaluation and processing of essential frames, thereby reducing delays in the processor and the Human-Machine Interface (HMI) warning module. Comprehensive validation is systematically conducted within a controlled test track environment to ensure precise execution of protocols, maintaining inputs closely aligned with real-time scenarios. The test methodology comprised the execution of pre-defined protocols that is steering robot and a technology-neutral procedure. Pre-defined protocols are scenarios created using the aforementioned assessing inputs. Cartesian coordinates of the system’s camera and driver eye point relative to the seating reference point (SgRP) are identified using a coordinate measurement machine (CMM) to measure the driver's position within the camera's field of view and mark the visibility zones. The protocols are executed with precision using a global navigation satellite system (GNSS), visual sensor, audio sensor and data logger. Subsequently, the system is tested with number of drivers trained on the Karolinska Sleepiness Scale (KSS) to conduct technology-neutral method for statistical analysis. Detailed analysis of the tested data, concluded with results and explored future prospects for quantifying driver drowsiness are discussed. The paper also discussed observations and challenges associated with the functionality of conventional systems and protocols currently deployed in the market.
Balasubrahmanyan, ChappagaddaAkbar Badusha, A
Edge Artificial Intelligence (AI) is poised to usher in a new era of innovations in automotive and mobility. In concert with the transition towards software-defined vehicle (SDV) architectures, the application of in-vehicle edge AI has the potential to extend well beyond ADAS and AV. Applications such as adaptive energy management, real-time powertrain calibration, predictive diagnostics, and tailored user experiences. By moving AI model execution right into edge, i.e. the vehicle, automakers can significantly reduce data transmission and processing costs, ensure privacy of user data, and ensure timely decision-making, even when connectivity is limited. However, achieving such use of edge AI will require essential cloud and in-vehicle infrastructure, such as automotive-specific MLOps toolchains, along with the proper SDV infrastructure. Elements such as flexible compute environments, deterministic and high-speed networks, seamless access to vehicle-wide data and control functions. This paper outlines examples of edge AI use cases beyond ADAS and AD, the challenges current vehicle electrical/electronic (E/E) architectures pose, and the limitations of general-purpose MLOps tools. It goes on to discuss how the shift to Software-Defined Vehicle (SDV) architectures and MLOps toolchains that are needed to overcome the challenges.
Khatri, SanjaySah, Mohamadali
With rapid advancements in Autonomous Driving (AD) & Advanced Driver Assistance Systems (ADAS), numerous sensors are integrated in vehicles to achieve higher and reliable level of autonomy. Due to the growing number of sensors and its fusion creates complex architecture which causes challenges in calibration, cost, and system reliability. Considering the need for further ADAS advancements and addressing the challenges, this paper evaluates a novel solution called One Radar - a single radar system with a wide field of view enabled by advanced antenna design. Placing the single radar at the rear of the vehicle eliminates the need for corner radars and ultrasonic sensors used for parking assistance. With rigorous real-world testing in different urban and low-speed scenarios, the single radar solution showed comparable accuracy in object detection with warning and parking assistance to the conventional combination of corner radars and ultrasonic sensors. The simple single sensor-based architecture not only reduces signal processing complexity and development time but also minimizes interference risks that comes with multiple sensor setup. This innovation results in a significant reduction in the Bill of Materials (BOM) for manufacturers by up to 40% for rear/side sensing modules, lowering production costs and enabling more affordable ADAS-equipped vehicles for end customers. Additionally, the simplified design enhances scalability for mass market adoption. The research paper talks about the single radar performance in various use cases to validate the features such as Rear parking alert system (RPAS), Door open warning (DOW), Blind spot detection (BSD), Lane change warning (LCW) and Rear collision warning (RCW) functionalities, highlighting its versatility as a standalone sensor module for future autonomous systems.
Anandan, RamSharma, Akash
Robust validation of Advanced Driver Assistance Systems (ADAS) considering real-world conditions is a vital for ensuring safety. Mileage accumulation is a one of the validation method for ensuring ADAS system robustness. By subjecting systems to diverse real-world driving environments and edge-case scenarios, engineers can evaluate performance, reliability, and safety under realistic conditions. In accordance with ISO 21448 (SOTIF), known hazardous scenarios are explicitly tested during robustness validation in combination of virtual and physical testing at component, sub system and vehicle level, while unknown hazards may emerge through extended mileage by running vehicles on roads, allowing them to be identified and classified. However, defining a mileage target that ensures comprehensive safety remains a significant engineering challenge. This paper proposes a data-driven approach to define mileage accumulation targets for validating Autonomous Emergency Braking Systems (AEBS), using detailed analysis of real-world accident data in India along with ISO 21448 (SOTIF) validation framework. National-level accident data from MoRTH and the RASSI database, along with statistics for medium and heavy commercial vehicles, are utilized to derive the base incident rate for frontal collisions that AEBS is intended to mitigate. The framework integrates critical factors such as hazardous behavior probability, controllability, and severity to calculate a target incident rate, which then informs the required test mileage needed to statistically validate AEBS performance at a specified confidence level. The study outlines the derivation of mileage requirements by considering both accident and fatality reduction as primary safety metrics. This approach provides engineering guidance for defining test mileage requirements with respective to the defined acceptance criteria that ensure AEBS system robust validation considering the real world scenarios in India.
Koralla, SivaprasadRavjani, AminTatikonda, VijayGadekar, Ganesh
The purpose of this report is to identify systematic approach of formation of India specific automotive database matrix. At first the paper reviews the practices used to prepare automotive dataset catalogue with established pattern to showcase automotive dataset from which appropriate data clusters can be picked up judiciously in order to train ADAS algorithms. The work applies this framework which helps to establish strategy to build a grid in which Indian automotive dataset can be contoured and selection of serviceable data bunches can be picked. This would make sure prompt selection of database aiming model training with valid input. This serves the purpose of implementation and evaluation of varied ADAS levels in India which insist upon good quality of distinguished dataset pertaining to Indian scenarios. The paper describes the approach with the example of AEB scenarios and present appropriate matrix readiness comprising of relevant data objects excluding unnecessary junk data targeting aftermath. The methodology can be base for amalgam of various Indian specific scenarios layered as per traffic objects and functions which can be picked up based on prerequisites of the test criteria and model training.
Behere, Sayali RajendraKarle, ManishKarle, Ujjwala
Road accidents involving cut-in and sudden brake events on highways present major challenges to driver safety, often outpacing the response time of traditional Advanced Driver Assistance Systems (ADAS). The objective of this study is to predict potential collisions caused by cut-ins before ADAS intervention becomes necessary, allowing for earlier driver alerts and enhanced vehicle response. The proposed method employs machine learning and deep learning approaches, specifically Long Short-Term Memory (LSTM) networks, to forecast collision risks 0.5 to 3 seconds in advance. Synthetic data generation techniques are used to create rare but critical cut-in and braking scenarios, complementing real-world data from test vehicles and accident records. Key predictive features monitored include relative velocity, lateral velocity, and lane overlap, which provide dynamic indicators of imminent risk. Results show that the system achieves an average early warning time of 1.35 seconds in 40.206% of evaluated hazardous scenarios, significantly improving the chance for evasive maneuvers and collision avoidance. Compared to conventional reactive systems, our approach proactively identifies threats by integrating real-time sensing with predictive modeling. The conclusion drawn from this research is that combining synthetic event generation with LSTM-based predictive analytics can substantially enhance ADAS capabilities, reduce accident rates, and pave the way for smarter, more anticipatory vehicle safety systems. These findings offer an important advancement toward more intelligent road safety technologies that emphasize prevention rather than reaction.
Srivastava, RohanNayak, Apoorva S.Suvvari, Sai DileepSatwik, RahulBhattacharya, Abhinov
Simulation has become mission-critical for ADAS development. Model-based systems engineering can integrate modeling and simulation from the start of the design process. Advanced Driver Assistance Systems (ADAS) are transforming vehicle safety, acting as the bridge between conventional driving and full autonomy. From adaptive cruise control to emergency braking and blind-spot detection, these technologies rely on a dense network of radar sensors, antennas, electronic control units and software. What unites them is the need for precise functionality under complex real-world situations. Achieving full reliability requires more than testing on the road; it demands a virtual approach grounded in simulation. Simulation has become mission-critical for ADAS development. As new vehicles integrate dozens of sensors into tightly constrained spaces, even subtle design decisions can affect system performance. Radar solutions, in particular, present unique challenges, especially as vehicle surfaces grow more complex and the number of onboard systems increases.
Eichler, Jan
This article suggests a validation methodology for autonomous driving. The goal is to validate front camera sensors in advanced driver-assist systems (ADAS) based on virtually generated scenarios. The outcome is the CARLA-based hardware-in-the-loop (HIL) simulation environment (CHASE). It allows the rapid prototyping and validation of the ADAS software. We tested this general approach on a specific experimental application/setup for a vehicle front camera sensor. The setup results were then proven to be comparable to real-world sensor performance. The CARLA simulation environment was used in tandem with a vehicle CAN bus interface. This introduced a significantly improved realism to user-defined test scenarios and their results. The approach benefits from almost unlimited variability of traffic scenarios and the cost-efficient generation of massive testing data.
Cardozo, Shawn MosesHlavác, Václav
Vehicle-to-whatever communication technologies continue to be put through their paces around the world. To point at just one example of the continued evolution of V2X technologies, let's take a quick visit to Japan and the 2025 JSAE Annual Spring Congress this May. That's where Toyota and Eye-Net Mobile Ltd., a subsidiary of Foresight Autonomous Holdings Ltd., presented a research paper on using vehicle-to-network technology to enhance ADAS systems by connecting to smartphones in the environment to address the inherent limitations of in-vehicle sensors. Titled, “Feasibility Study of a Hazard Avoidance Brake Control System Using V2N Technology,” the paper examined how smartphones could act as external sensors that could connect to an onboard ADAS system using vehicle-to-network (V2N) communications. The main purpose of these signals would be for enhanced hazard detection, Toyota said, adding that some of the top issues addressed in the paper were “communication latency, tracking accuracy, and the positioning precision of these devices in diverse urban environments,” where direct line-of-sight isn't always possible.
Blanco, Sebastian
While electric powertrains are driving 48V adoption, OEMs are realizing that xEV and ICE vehicles can benefit from a shift away from 12-volt architectures. In every corner of the automotive power engineering world, there are discussions and debates over the merits of 48V power networks vs. legacy 12V power networks. The dialogue started over 20 years ago, but now the tone is more serious. It's not a case of everything old is new again, but the result of a growing appetite for more electrical power in vehicles. Today's vehicles - and the coming generations - require more power for their ADAS and other safety systems, infotainment systems and overall passenger comfort systems. To satisfy the growing demand for low-voltage power, it is necessary to boost the capacity of the low-voltage power network by two or three times that of the late 20th century. Delivering power is more efficient at a higher voltage, and today, 48V is the consensus voltage for that higher level.
Green, Greg
This study presents a two-step method for estimating motorcycle tire lateral forces, which are critical to the safety of driver assistance systems. In the pre-filtering stage, a partial attitude of the motorcycle is estimated using a Kalman filter and a kinematic model. In the observation stage, the side slip angle and subsequently the tire lateral forces are provided by a sliding mode observer. It extends previous research by incorporating both out-of-plane and in-plane dynamics. The paper also proposes an approach for selecting the Kalman filter parameters. An approach to identify the stochastic sensor errors of the inertial measurement unit is presented. The identified parameters are used as a basis for the selection of the covariances. The overall study provides a practical implementation strategy and demonstrates its applicability in real-world scenarios. The experiments show the results of the lateral force estimation and its relation to the friction ellipse. The effectiveness of the proposed observer concept is evaluated using simulation data and measured data.
Winkler, AlexanderGrabmair, GernotReger, Johann
There are many riders who drive motorcycles on winding mountain roads and caused single motorcycle traffic accidents on curved roads by lane departure. Driving a motorcycle requires subtle balancing and maneuvering. In this study, in order to clarify the influence of lane departure caused by inadequate driving maneuvers against road alignment, the authors analyzed the required curve initial operation and driving maneuvers in curves depending on the traveling speed using a kinematics simulation for motorcycle dynamics. In addition, it was analyzed how inadequate driving maneuvers for curved roads can easily cause lane departure. As a result, it shows that the steering maneuvers and the lean of motorcycle body during the curves are highly affected by the vehicle speed, and the required maneuvers increases rapidly with increasing speed. The inadequate maneuver in the curves, especially for the lean of motorcycle body and steering torque, even by 10%, may cause failure to follow the correct driving path, which may lead to lane departure. Furthermore, the results suggest that road alignment with longitudinal slope and superelevation are more susceptible to the effects of variable steering torque factors, and that a slight steering wobble can easily lead to lane departure on Hilly and Mountainous Road (hereinafter called HMR). The verification with traffic accident statistical data, indicated that downhill left curves and uphill right curves are the contributing factors to the higher occurrence of fatal and serious injury motorcycle accidents. Therefore, safety measures for motorcycles need to be strengthened for these situations.
Kuniyuki, HiroshiTakechi, So
Precisely understanding the driving environment and determining the vehicle’s accurate position is crucial for a safe automated maneuver. vehicle following systems that offer higher energy efficiency by precisely following a lead vehicle, the relative position of the ego vehicle to lane center is a key measure to a safe automated speed and steering control. This article presents a novel Enhanced Lane Detection technique with centimeter-level accuracy in estimating the vehicle offset from the lane center using the front-facing camera. Leveraging state-of-the-art computer vision models, the Enhanced Lane Detection technique utilizes YOLOv8 image segmentation, trained on a diverse world driving scenarios dataset, to detect the driving lane. To measure the vehicle lateral offset, our model introduces a novel calibration method using nine reference markers aligned with the vehicle perspective and converts the lane offset from image coordinates to world measurements. This design minimizes the sensitivity of offset estimation to lane detection accuracy and vehicle orientation. Compared to the existing deep learning-based depth perception models and stereo vision systems, our calibration method significantly improves postprocessing time and minimizes the impacts of the processing delay on the vehicle following system energy efficiency. To assess the accuracy and processing time, we implemented the model on an instrumented L4-capable vehicle and conducted automated vehicle following tests in a controlled environment. In our tests, the model achieved a high level of accuracy, with a biased error of only 0.214 m and a random walk error standard deviation of 0.135 m, demonstrating its reliability across various environmental conditions and ensuring precise lane tracking. Results demonstrate reliable performance across various environmental conditions and sensor noise levels, ensuring precise lane tracking and enhanced automated maneuvering.
Karuppiah Loganathan, Nirmal RajaPoovalappil, AmanNaber, JeffreyRobinette, DarrellBahramgiri, Mojtaba
Adaptive Cruise Control (ACC) is an advanced driver assistance system designed to manage a vehicle's longitudinal motion. Its effectiveness is critically dependent on the precision of the sensors used. While ACC algorithms are optimized for performance, the overall efficacy of the system is significantly influenced by sensor accuracy and variability. Quantifying the impact of these factors on ACC performance poses a challenge. This paper explores the effects of sensor accuracy on ACC performance through a simulation study that replicates the sensor accuracy and variability observed in realworld vehicles. Additionally, the paper examines potential strategies to mitigate performance fluctuations caused by sensor variability.
Awathe, ArpitVarunjikar, TejasRaut, Abhinandan VijayPatel, Darsh
In the domain of advanced driver assistance systems and autonomous vehicles, precise perception and interpretation of the vehicle's environment are not merely requirements they are the very foundation upon which every aspect of functionality and safety is constructed. One prevalent method of representing the environment is through the use of an occupancy grid map. This map segments the environment into distinct grid cells, each of which is evaluated to determine if it is occupied or free. This evaluation operates under the assumption that each grid cell is independent of the others. The underlying mathematical structure of this system is the binary Bayes filter (BBF). The BBF integrates sensor data from various sources and can incorporate measurements taken at different times. The occupancy grid map does not rely on the identification of individual objects, which allows it to depict obstacles of any shape. This flexibility is a key advantage of this approach. Traditional occupancy grid maps fall short when it comes to predicting dynamic environments due to their lack of a process model. A notable enhancement to this static model is the Bayesian Occupancy Filter (BOF), which, unlike its predecessor, estimates a velocity distribution for each grid cell's occupancy using a histogram filter. However, the BOF's computational demands are high. To address this, research propose representing the dynamic state of grid cells using particles. This method enables the computation of dynamic grid maps in real-time applications, even with larger grid cell sizes and higher resolution. Despite these advancements, dynamic occupancy grid maps remain a relatively new field of study, especially when compared to more established object-tracking approaches. Until now, the BOF has been treated as a distinct research area with minimal overlap with other tracking methodologies. This methodology aims to bridge that gap and foster a more integrated approach to dynamic environment estimation. This study introduces a novel approach to dynamic grid mapping, conceptualized as an approximation of a Random Finite Set (RFS) filter. An RFS is a probabilistic representation of a finite, random collection of objects and their respective states. Finite Set Statistics (FISST) provide a framework for Bayesian filtering of these random finite sets and form the foundation for several multi-object tracking methodologies, such as the Probability Hypothesis Density (PHD) filter. By characterizing the grid as an RFS, we can apply sophisticated concepts from the well-established domain of RFS filtering to dynamic grid mapping. The research develops a filter known as the Probability Hypothesis Density/Multi-Instance Bernoulli (PHD/MIB) filter. This filter alternately represents and propagates the dynamic grid map as a PHD and as multiple instances of Bernoulli filters, thereby offering a more integrated and efficient approach to dynamic environment estimation. Furthermore, this research introduces a Sequential Monte Carlo (SMC) implementation of the PHD/MIB filter, as well as an approximation within the Dempster-Shafer framework, termed the Dempster-Shafer PHD/MIB (DS-PHD/MIB) filter. This DS-PHD/MIB filter necessitates fewer particles than the original PHD/MIB filter, thereby enhancing computational efficiency. The study provides a comprehensive description of an efficient, massively parallel implementation of the DS-PHD/MIB filter. The algorithm's pseudo code is also outlined, offering a clear and concise understanding of its workings. This approach further strengthens the integrated and efficient methodology for dynamic environment estimation. In conclusion, the research delineates the attributes of the DS-PHD/MIB filter and debates its pros and cons in comparison to object-based tracking methodologies, using practical examples for illustration. A quantitative assessment using real-world data demonstrates that the DS-PHD/MIB filter yields consistent state estimation outcomes. It effectively models both the stochastic multi-object transition process and the stochastic multi-object observation process. Moreover, the evaluation affirms the real-time capability of the parallelized implementation of the DS-PHD/MIB filter. It validates its utility for state estimation in dynamic vehicle environments, thereby underscoring its potential for practical applications in dynamic environment estimation. This comprehensive approach offers a promising avenue for future research and development in this field.
Wani, AnkitIthape, AvinashSingh, JyotsanaBurangi, PiyushBorawar, Amit
Deliberate modifications to infrastructure can significantly enhance machine vision recognition of road sections designed for Vulnerable Road Users, such as green bike lanes. This study evaluates how green bike lanes, compared to unpainted lanes, enhance machine vision recognition and vulnerable road users safety by keeping vehicles at a safe distance and preventing encroachment into designated bike lanes. Conducted at the American Center for Mobility, this study utilizes a vehicle equipped with a front-facing camera to assess green bike lane recognition capabilities across various environmental conditions including dry daytime, dry nighttime, rain, fog, and snow. Data collection involved gathering a comprehensive dataset under diverse conditions and generating masks for lane markings to perform comparative analysis for training Advanced Driver Assistance Systems. Quality measurement and statistical analysis are used to evaluate the effectiveness of machine vision recognition using metrics, such as Blind/Reference-less Image Spatial Quality Evaluator, Naturalness Image Quality Evaluator, and Entropy-based Image Quality Assessment. The results indicate that green bike lanes are more likely to be recognized by machine vision systems across a wide range of environmental conditions, demonstrating enhanced recognition capabilities. Green lane markings exhibit enhanced visibility and stability, with BRISQUE scores below 82, a median contrast ratio of 17.6, and improved resilience to motion blur and NIQE variations under diverse conditions.
Ponnuru, Venkata Naga RithikaDas, SushantaGrant, JosephNaber, JeffreyBahramgiri, Mojtaba
Personalization is a growing topic in the automotive space, where Artificial Intelligence can be used to deliver a customized experience in features like seat positioning and climate control. Considering that the leading cause of accidents is driving at an inappropriate speed, personalizing the speed limit for a driver can greatly improve vehicle safety. Current speed limits apply to all drivers, irrespective of skill, including special speed limits when there are adverse weather conditions. As these speed limits do not consider an individual’s skill and capabilities, the limit could still be inappropriate for a given driver in that specific driving context. Therefore, we propose a system that can profile the driver’s style to recommend a personalized speed limit, based on both the environmental context and their skill in that environment. The system uses a neural network to classify the driver’s behavior in specific environments by monitoring the vehicle data and the environmental conditions. The network is trained to identify situations when the driver is not comfortable in the current driving context and recommends a more appropriate speed limit. This personalized limit is based on traffic sign speed limits, weather conditions, lighting conditions, and the driver profile. Personalization in this feature could reduce the probability of accidents, especially when the driver is safer driving at speeds lower than the road speed limit based on the weather conditions.
Perumal, RathapriyaChouhan, MadhvendraRangarajan, Rishi
Items per page:
1 – 50 of 850