Browse Topic: Autonomous vehicles

Items (2,807)
Towards the goal of real-time navigation of autonomous robots, the Iterative Closest Point (ICP) based LiDAR odometry methods are a favorable class of Simultaneous Localization and Mapping (SLAM) algorithms for their robustness under any light conditions. However, even with the recent methods, the traditional SLAM challenges persist, where odometry drifts under adversarial conditions such as featureless or dynamic environments, as well as high motion of the robots. In this paper, we present a motion-aware continuous-time LiDAR-inertial SLAM framework. We introduce an efficient EKF-ICP sensor fusion solution by loosely coupling poses from the continuous time ICP and IMU data, designed to improve convergence speed and robustness over existing methods while incorporating a sophisticated motion constraint to maintain accurate localization during rapid motion changes. Our framework is evaluated on the KITTI datasets and artificially motion-induced dataset sequences, demonstrating
Kokenoz, CigdemShaik, ToukheerSharma, AbhishekPisu, PierluigiLi, Bing
The advent of autonomous vehicles (AVs) marks a revolutionizing transformation in transportation, with the potential to significantly enhance safety and efficiency through advanced trajectory planning and optimization capabilities. A crucial component in realizing these benefits is the use of optimization-based control strategies for real-time path planning. Among these, model predictive path integral (MPPI) control algorithms stand out as a sampling-based stochastic control method, offering precise control in dynamic environments through random sampling. While the MPPI control has shown promising results, there has been limited investigation into the effects of different prediction horizon times on control performance of these algorithms. This paper seeks to address this gap by proposing a multi-input MPPI control method for AVs using a single-track vehicle dynamic model. Our research focuses on the influence of various prediction horizon times on trajectory optimization during lane
Yang, YanwenNegash, NatnaelYang, James
The rapid development of autonomous vehicles necessitates rigorous testing under diverse environmental conditions to ensure their reliability and safety. One of the most challenging scenarios for both human and machine vision is navigating through rain. This study introduces the Digitrans Rain Testbed, an innovative outdoor rain facility specifically designed to test and evaluate automotive sensors under realistic and controlled rain conditions. The rain plant features a wetted area of 600 square meters and a sprinkled rain volume of 600 cubic meters, providing a comprehensive environment to rigorously assess the performance of autonomous vehicle sensors. Rain poses a significant challenge due to the complex interaction of light with raindrops, leading to phenomena such as scattering, absorption, and reflection, which can severely impair sensor performance. Our facility replicates various rain intensities and conditions, enabling comprehensive testing of Radar, Lidar, and Camera
Feichtinger, Christoph Simon
Connected and automated vehicle (CAV) technology is a rapidly growing area of research as more automakers strive towards safer and greener roads through its adoption. The addition of sensor suites and vehicle-to-everything (V2X) connectivity gives CAVs an edge on predicting lead vehicle and connected intersection states, allowing them to adjust trajectory and make more fuel-efficient decisions. Optimizing the energy consumption of longitudinal control strategies is a key area of research in the CAV field as a mechanism to reduce the overall energy consumption of vehicles on the road. One such CAV feature is autonomous intersection navigation (AIN) with eco-approach and departure through signalized intersections using vehicle-to-infrastructure (V2I) connectivity. Much existing work on AIN has been tested using model-in-loop (MIL) simulation due to being safer and more accessible than on-vehicle options. To fully validate the functionality and performance of the feature, additional
Hamilton, KaylaMisra, PriyashrabaOrd, DavidGoberville, NickCrain, TrevorMarwadi, Shreekant
Technology development for enhancing passenger experience has gained attention in the field of autonomous vehicle (AV) development. A new possibility for occupants of AVs is performing productive tasks as they are relieved from the task of driving. However, passengers who execute non-driving-related tasks are more prone to experiencing motion sickness (MS). To understand the factors that cause MS, a tool that can predict the occurrence and intensity of MS can be advantageous. However, there is currently a lack of computational tools that predict passenger's MS state. Furthermore, the lack of real-time physiological data from vehicle occupants limits the types of sensory data that can be used for estimation under realistic implementations. To address this, a computational model was developed to predict the MS score for passengers in real time solely based on the vehicle's dynamic state. The model leverages self-reported MS scores and vehicle dynamics time series data from a previous
Kolachalama, SrikanthSousa Schulman, DanielKerr, BradleyYin, SiyuanWachsman, Michael BenPienkny, Jedidiah Ethan ShapiroJalgaonkar, Nishant M.Awtar, Shorya
Trajectory tracking control is a key component of vehicle autonomous driving technology. Compared with traditional vehicles, Distributed Driven Electric Vehicle (DDEV) is an ideal vehicle for trajectory tracking control because of its high space utilization, redundant control freedom and fast system response. However, the chassis execution system of DDEV has a relatively large number of sensors, which significantly increases its probability of failure. In this paper, we propose a trajectory tracking fault-tolerant control method for DDEV considering steering actuator faults. Firstly, we establish the dynamic model of the steering actuator and the trajectory tracking model of DDEV. The model is linearized and discretized by using Taylor series expansion and forward Euler method. Next, considering multi-objective constraints such as motion comfort, actuator saturation and road adhesion boundary, the trajectory tracking control strategy of DDEV is designed by using model predictive
Wang, DepingLi, LunTeng, YuhanZhu, BingChen, Zhicheng
LiDAR sensors have become an integral component in the realm of autonomous driving, widely utilized in environmental perception and vehicle navigation. However, in real-world road environments, contaminants such as dust and dirt can severely hamper the cleanliness of LiDAR optical windows, thereby degrading operational performance and affecting the overall environmental perception capabilities of intelligent driving systems. Consequently, maintaining the cleanliness of LiDAR optical windows is crucial for sustaining device performance. Unfortunately, the scarcity of publicly available LiDAR contamination datasets poses a challenge to the research and development of contamination identification algorithms. This paper first introduces a method for acquiring LiDAR-pollution datasets. LiDAR data acquisition on urban open roads simulates different types of pollution, including mud and leaves. The constructed dataset meticulously differentiates among the three states with clear labels: no
Wei, ZiyuQuo, BinyunLujia, RanLi, Liguang
Path tracking control, which is one of the most important foundations of autonomous driving, could help the vehicle to precisely and smoothly follow the preset path by actively adjusting the front wheel steering angle. Although there are a number of advanced control methods with simple structure and reliable robustness that could assist vehicles achieving path tracking, these controllers have many parameters to be calibrated, and there is a lack of guidance documents to help non-professional test site engineers quickly master calibration methods. Therefore, this paper proposes a parameter virtual calibration method based on the deep reinforcement learning, which provides an effective solution for parameter calibration of vehicle path tracking controller. Firstly, the vehicle trajectory tracking model is established through the kinematic relationship between the vehicle and the target path, combined with the Taylor series expansion linearization method. Next, a vehicle path tracking
Zhao, JianGuo, ChenghaoZhao, HuiChaoZhao, YongqiangYu, ZhenZhu, BingChen, Zhicheng
With the continuous advancement of artificial intelligence technology, the automation level of electric vehicles (EVs) is rapidly increasing. Despite the improvements in travel efficiency, safety, and convenience brought about by automation, cutting-edge intelligent technologies also pose the potential of increased energy consumption, such as the computational power required by advanced algorithms and the energy usage of high-precision equipment, leading to higher overall energy consumption for connected or autonomous electric vehicles (CAEVs). To assess the impact of intelligent technologies on AEVs, this study innovatively provides a comprehensive evaluation of the impact of intelligent technologies on CAEV energy consumption from both positive and negative perspectives. After reviewing 59 relevant studies, the findings highlight energy savings achieved through Vehicle-to-Infrastructure and Vehicle-to-Vehicle cooperation as positive effects, while increased energy consumption from
Liu, TianyiQi, HaoOu, Shiqi (Shawn)
The shift towards hybrid and electric powertrains in off-road vehicles aims to enhance mobility, extend range, and improve energy efficiency. However, heat pump-based battery thermal management systems in these vehicles continue to consume significant energy, impacting overall range and efficiency. Effective thermal management is essential for maintaining battery performance and safety, particularly in extreme conditions. Although high-fidelity models can capture the complex dynamics of heat pumps, real-time control within model-based optimization frameworks often depends on simplified models, which can degrade system performance. To address this, we propose a novel data-driven grey box control-oriented model (COM) that accurately represents the thermal dynamics of a vapor-compression refrigeration-based heat pump system. This COM is integrated into a model-predictive control (MPC) framework, optimizing thermal management during transient and burst-power operations of the battery pack
Sundar, AnirudhGhate, AtharvaZhu, QilunPrucka, RobertRuan, YeefengFigueroa-Santos, MiriamBarron, Morgan
Light Detection and Ranging (LiDAR) is a promising type of sensor for autonomous driving that utilizes laser technology to provide perceptions and accurate distance measurements of obstacles in the vehicle path. In recent years, there has also been a rise in the implementation of LiDARs in modern and autonomous vehicles to aid self-driving features. However, navigating adverse weather remains one of the biggest challenges in achieving Level 5 full autonomy due to sensor soiling, leading to performance degradation that can pose safety hazards. When driving in rain, raindrops impact the LiDAR sensor assembly and cause attenuation of signals when the light beams undergo reflections and refractions. Consequently, signal detectability, accuracy, and intensity are significantly affected. To date, limited studies have been able to perform objective evaluations of LiDAR performance, most of which faced limitations that hindered realistic, controllable, and repeatable testing. Therefore, this
Pao, Wing YiLi, LongAgelin-Chaab, MartinRoy, LangisKnutzen, JulianBaltazar, AlexisMuenker, KlausChakraborty, AnirbanKomar, John
The slope and curvature of spiral ramps in underground parking garages change continuously, and often lacks of predefined map information. Traditional planning algorithms is difficult to ensure safety and real-time performance for autonomous vehicles entering and exiting underground parking garages. Therefore, this study proposed the Model Predictive Path Integral (MPPI) method, focusing on solving motion planning problems in underground parking garages without predefined map information. This sample-based method to allows simultaneous online autonomous vehicle planning and tracking while not relying on predefined map information,along with adjusting the driving path accordingly. Key path points in the spiral ramp environment were defined by curvature, where reducing the dimensionality of the sampling space and optimizing the computational efficiency of sampled trajectories within the MPPI framework. This ensured the safety and computational speed of the improved MPPI method in motion
Liu, ZuyangShen, YanhuaWang, Kaidi
As automotive technology advances, the need for comprehensive environmental awareness becomes increasingly critical for vehicle safety and efficiency. This study introduces a novel integrated wind, weather, and motion sensor designed for moving objects, with a focus on automotive applications. The sensor’s potential to enhance vehicle performance by providing real-time data on local atmospheric conditions is investigated. The research employs a combination of sensor design, vehicle integration, and field-testing methodologies. Findings prove the sensor’s capability to accurately capture dynamic environmental parameters, including wind speed and direction, temperature, and humidity. The integration of this sensor system shows promise in improving vehicle stability, optimizing fuel efficiency through adaptive aerodynamics, and enhancing the performance of autonomous driving systems. Furthermore, the study explores the potential of this technology in contributing to connected vehicle
Feichtinger, Christoph Simon
Autonomous ground navigation has advanced significantly in urban and structured environments, supported by the availability of comprehensive datasets. However, navigating complex and off-road terrains remains challenging due to limited datasets, diverse terrain types, adverse environmental conditions, and sensor limitations affecting vehicle perception. This study presents a comprehensive review of off-road datasets, integrating their applications with sensor technologies and terrain traversability analysis methods. It identifies critical gaps, including class imbalances, sensor performance under adverse conditions, and limitations in existing traversability estimation approaches. Key contributions include a novel classification of off-road datasets based on annotation methods, providing insights into scalability and applicability across diverse terrains. The study also evaluates sensor technologies under adverse conditions and proposes strategies for incorporating event-based and
Musau, HannahRuganuza, DenisIndah, DebbieMukwaya, ArthurGyimah, Nana KankamPatil, AshishBhosale, MayureshGupta, PrakharMwakalonge, JudithJia, YunyiMikulski, DariuszGrabowsky, DavidHong, Jae DongSiuhi, Saidi
As a crucial tool for lunar exploration, lunar rovers are highly susceptible to instability due to the rugged lunar terrain, making control of driving stability essential during operation. This study focuses on a six-wheel lunar rover and develops a torque distribution strategy to improve the handling stability of the lunar rover. Based on a layered control structure, firstly, the approach establishes a two-degree-of-freedom single-track model with front and rear axle steering at the state reference layer to compute the desired yaw rate and mass center sideslip angle. Secondly, in the desired torque decision layer, a sliding mode control-based strategy is used to calculate the desired total driving torque. Thirdly, in the torque distribution layer, the optimal control distribution is adopted to carry out two initial distributions and redistribution of the drive torque planned by the upper layer, to improve the yaw stability of the six-wheeled lunar rover. Finally, a multi-body dynamics
Liu, PengchengZhang, KaidiShi, JunweiYang, WenmiaoZhang, YunqingWu, Jinglai
As the electrification of chassis systems accelerates, the demand for fail-safety strategies is increasing. In the past, the steering system was mechanically connected, so the driver could respond directly to some extent. However, the Steer-by-Wire (SbW) system is composed of the column and rack bar as electrical signals, so the importance of response strategies for steering system failure is gradually increasing. When a steering system failure occurs, a differential braking control using the difference in braking force between the left and right wheels was studied. Recently, some studies have been conducted to model the wheel reaction force generated during a differential braking. Since actual tires and road surfaces are nonlinear and cause large model errors, model-based control methods have limited performance. Also, in previous studies assumed that the driver normally operates the steering wheel in a failure situation. However, if limited to a situation such as autonomous driving
Kim, SukwonKim, Young GwangKim, SungDoMoon, Sung Jin
The unicycle self-balancing mobility system offers superior maneuverability and flexibility due to its unique single-wheel grounding feature, which allows it to autonomously perform exploration and delivery tasks in narrow and rough terrains. In this paper, a unicycle self-balancing robot traveling on the lunar terrain is proposed for autonomous exploration on the lunar surface. First, a multi-body dynamics model of the robot is derived based on quasi-Hamilton equations. A three-dimensional terramechancis model is used to describe the interaction between the robot wheels and the lunar soil. To achieve stable control of the robot's attitude, series PID controllers are used for pitch and roll attitude self-balancing control as well as velocity control. The whole robot model and control strategy were built in MATLAB and the robot's traveling stability was analyzed on the lunar terrain.
Shi, JunweiZhang, KaidiDuan, YupengWu, JinglaiZhang, Yunqing
Since the introduction of ABS (1978), TCS (1986) and ESC (1995) in series production, the number of modern vehicle dynamics control functions and advanced driver assistance systems (ADAS) has been continuously increasing. Meanwhile, many functions are available that influence vehicle motion (vehicle dynamics). Since these are only partially and not hierarchically coordinated, the control of vehicle motion is still suboptimal. Current megatrends (automated driving, electromobility, software-defined vehicles) and new key technologies (steer-by-wire, brake-by-wire, domain-based E/E architectures) lead to an increasing number of electrified, motion-relevant components being introduced into series production. These components enable the development of an integrated chassis control (ICC) that controls all motion-relevant components, networks them with each other and coordinates them holistically to optimally control the vehicle motion regarding an adjustable desired driving behavior. Vehicle
Wielitzka, MarkAhrenhold, TimVocht, MoritzRawitzer, JonasSchrader, Jonas
Commercial Vehicle (CV) market is growing rapidly with the advancement of Software-Defined Vehicles (SDVs), which provide greater level of flexibility, efficiency and integration of AI & cutting-edge technology. This research provides an in-depth analysis of E&E architecture of CVs, focusing on the integration of SDV-based technology, which represents the transition from hardware-focused to a more dynamic, software-focused methodology. The research begins with the fundamental concepts of E&E architecture in CVs, including virtualization, centralized computing, feature based ECU, CAN and modular frameworks which are then upgraded to meet various operational and customer requirements. The capacity of SDV-based architecture designs to scale to handle heavy duty commercial vehicles is a primary focus, with an emphasis on ensuring the safety and security, to defend against potential vulnerabilities. Furthermore, the integration of real-time data processing capabilities and advanced E&E
Saini, VaibhavJain, AyushiMeduri, PramodaSolutions GmbH, Verolt Technology
One challenge for autonomous vehicle (AV) control is the variation in road roughness which can lead to deviations from the intended course or loss of road contact while steering. The aim of this work is to develop a real-time road roughness estimation system using a Bayesian-based calibration routine that takes in axle accelerations from the vehicle and predicts the current road roughness of the terrain. The Bayesian-based calibration method has the advantage of providing posterior distributions and thus giving a quantifiable estimate of the confidence in the prediction that can be used to adjust the control algorithm based on desired risk posture. Within the calibration routine, a Gaussian process model is first used as a surrogate for a simulated half-vehicle model which takes vehicle velocity and road surface roughness (GD) to output the axle acceleration. Then the calibration step takes in the observed axle acceleration and vehicle velocity and calibrates the Gaussian process model
Lewis, EdwinaParameshwaran, AdityaRedmond, LauraWang, Yue
This study introduces an innovative torque vectoring control strategy designed to enhance ride comfort in autonomous electric vehicles. The approach seamlessly integrates steering and rear axle force control within a model predictive control (MPC) framework, enabling real-time optimization of comfort and handling performance. The proposed control method is applied to a two-rear-motor vehicle model, where the MPC algorithm adjusts steering angles and tire forces to minimize discomfort caused by yaw rate and lateral acceleration. Simulation results from a lane-change scenario demonstrate significant improvements in comfort metrics compared to conventional torque vectoring control strategies. The findings highlight the ability of the proposed method to significantly enhance ride comfort without compromising vehicle dynamics. This integrated and adaptive control strategy offers a promising solution for improving passenger satisfaction in autonomous electric vehicles, with potential
Zhao, BolinLou, BaichuanHe, XianqiXue, WanyingLv, Chen
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
Fernandez, DavidMohajerAnsari, PedramSalarpour, AmirPesé, Mert D.
Advanced driver assistance systems (ADASs) and driving automation system technologies have significantly increased the demand for research on vehicle-state recognition. However, despite its critical importance in ensuring accurate vehicle-state recognition, research on road-surface classification remains underdeveloped. Accurate road-surface classification and recognition would enable control systems to enhance decision-making robustness by cross-validating data from various sensors. Therefore, road-surface classification is an essential component of autonomous driving technologies. This paper proposes the use of tire–pavement interaction noise (TPIN) as a data source for road-surface classification. Traditional approaches predominantly rely on accelerometers and visual sensors. However, accelerometer signals have inherent limitations because they capture only surface profile properties and are often distorted by the resonant characteristics of the vehicle structure. Similarly, image
Yoon, YoungsamKim, HyungjooLee, Sang KwonLee, JaekilHwang, SungukKu, Sehwan
The application of multi-sensor fusion for enhanced distance estimation accuracy in dynamic environments is crucial for advanced driver assistance systems (ADAS) and autonomous vehicles. The Limitations of single sensors such as cameras or radar in adverse conditions motivate the use of combined camera and radar data to improve reliability, adaptability, and object recognition. A multi-sensor fusion approach using an Extended Kalman Filter (EKF) is proposed to combine sensor measurements with a dynamic system model, achieving robust and accurate distance estimation. The research utilizes the Mississippi State University Autonomous Vehicular Simulator (MAVS), a physics-based simulation platform, to generate realistic synthetic datasets incorporating sensor imperfections such as noise and missed detections to create a controlled environment for data collection. Data analysis is performed using MATLAB. Qualitative metrics such as visualization of fused data vs ground truth and
Ebu, Iffat AraIslam, FahmidaRafi, Mohammad AbdusShahidRahman, MahfuzurIqbal, UmarBall, John
Understanding the formation and behaviour of sprays and aerosols generated by vehicles traveling on wet surfaces is crucial due to their impact on vehicle soiling, visibility, and autonomous driving. These sprays and aerosols can reduce visibility for other drivers, contribute to traffic accidents, and reduce the operational capabilities of sensors for driving assistance systems and future autonomous vehicles. Despite the critical importance of understanding the physical properties of these sprays and aerosols for the testing and validation of sensors used in environmental perception and recognition, field data on this subject remains limited. The formation and behaviour of these sprays and aerosols are complex. A fraction of the trailing droplets and ligaments originates directly from the tyres, while the remainder is generated upon the impact of the particles ejected from the tyres with the vehicle’s wheel houses and other surfaces, resulting in either coalescence or further
Otxoterena, PaulKallhammer, Jan-ErikEriksson, PeterRonelov, Erik
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
Zhang, HaosenLi, YihangWu, Guangqiang
Less costs and higher efficiency may be constant technological pursuit. Despite the great success, data-driven AI development still requires multiple stages such as data collection, cleaning, annotation, training, and deployment to work together. We expect an end-to-end style development process that can integrate these processes, achieving an automatic data production and algorithm development process that can work with just clicks of the mouse. For this purpose, we explore an end-to-end style parking algorithm development pipeline based on procedural parking scenario synthetic data generation. Our approach allows for the automated generation of parking scenarios according to input parameters, such as scene construction, static and dynamic obstacles arrangement, material textures modification, and background changes. It then combines with the ego-vehicle trajectories into the scenarios to render high-quality images and corresponding label data based on Blender software. Utilizing
Li, JianWang, HanchaoZhang, SongMeng, ChaoRui, Zhang
Trailer parking is a challenging task due to the unstable nature of the vehicle-trailer system in reverse motion and the unintuitive steering actions required at the vehicle to accomplish the parking maneuver. This paper presents a strategy to tackle this kind of maneuver with an advisory graphic aid to help the human driver with the task of manually backing up the vehicle-trailer system. A kinematic vehicle-trailer model is derived to describe the low-speed motion of the vehicle-trailer system, and its inverse kinematics is established by generating an equivalent virtual trailer axle steering command. The advisory system graphics is generated based on the inverse kinematics and displays the expected trailer orientation given the current vehicle steer angle and configuration (hitch angle). Simulation study and animation are set up to test the efficacy of the approach, where the user can select both vehicle speed and vehicle steering angle freely, which allows the user to stop the
Cao, XinchengChen, HaochongAksun Guvenc, BilinGuvenc, LeventLink, BrianHarber, JohnRichmond, PeterFan, ShihongYim, Dokyung
Intelligent transportation systems and connected and automated vehicles (CAVs) are advancing rapidly, though not yet fully widespread. Consequently, traditional human-driven vehicles (HDVs), CAVs, and human-driven connected and automated vehicles (HD-CAVs) will coexist on roads for the foreseeable future. Simultaneously, car-following behaviors in equilibrium and discretionary lane-changing behaviors make up the most common highway operations, which seriously affect traffic stability, efficiency and safety. Therefore, it’s necessary to analyze the impact of CAV technologies on both longitudinal and lateral performance of heterogeneous traffic flow. This paper extends longitudinal car-following models based on the intelligent driver model and lateral lane-changing models using the quintic polynomial curve to account for different vehicle types, considering human factors and cooperative adaptive cruise control. Then, this paper incorporates CAV penetration rates, shared autonomy rates
Wang, TianyiGuo, QiyuanHe, ChongLi, HaoXu, YimingWang, YangyangJiao, Junfeng
Lane-keeping is critical for SAE Level 3+ autonomous vehicles, requiring rigorous validation and end-to-end interpretability. All recently U.S.-approved level 3 vehicles are equipped with lidar, likely for accelerating active safety. Lidar offers direct distance measurements, allowing rule-based algorithms compared to camera-based methods, which rely on statistical methods for perception. Furthermore, lidar can support a more comprehensive and detailed approach to studying lane-keeping. This paper proposes a module perceiving oncoming vehicle behavior, as part of a larger behavior-tree structure for adaptive lane-keeping using data from a lidar sensor. The complete behavior tree would include road curvature, speed limits, road types (rural, urban, interstate), and the proximity of objects or humans to lane markings. It also accounts for the lane-keeping behavior, type of adjacent and opposing vehicles, lane occlusion, and weather conditions. The algorithm was evaluated using
Soloiu, ValentinMehrzed, ShaenKroeger, LukePierce, KodySutton, TimothyLange, Robin
In cold and snowy areas, low-friction and non-uniform road surfaces make vehicle control complex. Manually driving a car becomes a labor-intensive process with higher risks. To explore the upper limits of vehicle motion on snow and ice, we use an existing aggressive autonomous algorithm as a testing tool. We built our 1:5 scaled test platform and proposed an RGBA-based cost map generation method to generate cost maps from either recorded GPS waypoints or manually designed waypoints. From the test results, the AutoRally software can be used on our test platform, which has the same wheelbase but different weights and actuators. Due to the different platforms, the maximum speed that the vehicle can reach is reduced by 1.38% and 2.26% at 6.0 m/s and 8.5 m/s target speeds. When tested on snow and ice surfaces, compared to the max speed on dirt (7.51 m/s), the maximum speed decreased by 48% and 53.9%, respectively. In addition to the significant performance degradation on snow and ice, the
Yang, YimingBos, Jeremy P.
Off-road vehicles are required to traverse a variety of pavement environments, including asphalt roads, dirt roads, sandy terrains, snowy landscapes, rocky paths, brick roads, and gravel roads, over extended periods while maintaining stable motion. Consequently, the precise identification of pavement types, road unevenness, and other environmental information is crucial for intelligent decision-making and planning, as well as for assessing traversability risks in the autonomous driving functions of off-road vehicles. Compared to traditional perception solutions such as LiDAR and monocular cameras, stereo vision offers advantages like a simple structure, wide field of view, and robust spatial perception. However, its accuracy and computational cost in estimating complex off-road terrain environments still require further optimization. To address this challenge, this paper proposes a terrain environment estimating method for off-road vehicle anticipated driving area based on stereo
Zhao, JianZhang, XutongHou, JieChen, ZhigangZheng, WenboGao, ShangZhu, BingChen, Zhicheng
With the surge in adoption of artificial intelligence (AI) in automotive systems, especially Advanced Driver Assistance Systems (ADAS) and autonomous vehicles (AV), comes an increase of AI-related incidents–several of which have ended in injuries and fatalities. These incidents all share a common deficiency: insufficient coverage towards safety, ethical, and/or legal requirements. Responsible AI (RAI) is an approach to developing AI-enabled systems that systematically take such requirements into account. Existing published international standards like ISO 21448:2022 (Safety of the Intended Functionality) and ISO 26262:2018 (Road Vehicles – Functional Safety) do offer some guidance in this regard but are far from being sufficient. Therefore, several technical standards are emerging concurrently to address various RAI-related challenges, including but not limited to ISO 8800 for the integration of AI in automotive systems, ISO/IEC TR 5469:2024 for the integration of AI in functional
Nelson, JodyLin, Christopher
The development of autonomous driving technology will liberate the space in the car and bring more possibilities of comfortable and diverse sitting postures to passengers, but the collision safety problem cannot be ignored. The aim of this study is to investigate the changes of injury pattern and loading mechanism of occupants under various reclined postures. A highly rotatable rigid seat and an integrated three-point seat belt were used, with a 23g, 50kph input pulse. Firstly, the sled test and simulation using THOR-AV in a reclined posture were conducted, and the sled model was verified effective. Based on the sled model, the latest human body model, THUMS v7, was used for collision simulation. By changing the angle of seatback and seat pan, 5 seat configurations were designed. Through the calculation of the volunteers' pose regression function, the initial position of THUMS body parts in different seat configurations was determined. The responses of human body parts were output
Yang, XiaotingWang, QiangLiu, YuFei, JingWang, PeifengLi, ZhenBai, Zhonghao
This paper explores the integration of two deep learning models that are currently being used for object detection, specifically Mask R-CNN and YOLOX, for two distinct driving environments: urban cityscapes and highway settings. The hypothesis underlying this work is that different methods of object detection will work best in different driving environments, due to the differences in their unique strengths as well as the key differences in those driving environments. Some of these differences in the driving environment include varying traffic densities, diverse object classes, and differing scene complexities, including specific differences such as the types of signs present, the presence or absence of stoplights, and the limited-access nature of highways as compared to city streets. As part of this work, a scene classifier has also been developed to categorize the driving context into the two categories of highway and urban driving, in order to allow the overall object detection
Patel, KrunalPeters, Diane
This paper presents a comparative study between many control techniques to investigate the efficiency of the path tracking in various driving scenarios. In this study the Model predictive control (MPC), the adaptive model predictive control (AMPC) and the Stanley controller are employed to ensure that the vehicle follows reference paths accurately and robustly under varying environmental and vehicular conditions. Two driving scenarios are utilized S-road and Curved-road with MATLAB/Simulink under three different vehicle speeds to investigate vehicle performance employing the root mean square error (RMSE) as the performance evaluation function. Particle swarm optimization (PSO) utilized for optimizing the six parameters of the MPC prediction horizon (P), Control horizon(m), manipulated variable rates, manipulated variables weights and two output variables weights. Four objective functions are employed with PSO and compared to each other in terms of the time domain regarding the RMSE of
Eldesouky, Dina M.MustafaAbdelaziz, Taha HelmyMohamed, Amr.E
Vehicle sideslip is a valuable measurement for ground vehicles in both passenger vehicle and racing contexts. At relevant speeds, the total vehicle sideslip, beta, can help drivers and engineers know how close to the limits of yaw stability a vehicle is during the driving maneuver. For production vehicles or racing contexts, this measurement can trigger Electronic Stability Control (ESC). For racing contexts, the method can be used for driver training to compare driver techniques and vehicle cornering performance. In a fleet context with Connected and Autonomous Vehicles (CAVS) any vehicle telemetry reporting large vehicle sideslip can indicate an emergency scenario. Traditionally, sideslip estimation methods involve expensive and complex sensors, often including precise inertial measurement units (IMUs) and dead reckoning, plus complicated sensor fusion techniques. Standard GPS measurements can provide Course Over Ground (COG) with quite high accuracy and, surprisingly, the most
Hannah, AndrewCompere, Marc
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, MeredithNguyen, AnHelber, NicholasHeydinger, Gary
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
Karuppiah Loganathan, Nirmal RajaPoovalappil, AmanNaber, JeffreyRobinette, DarrellBahramgiri, Mojtaba
This paper examines the challenges and mechanisms for ensuring Freedom from Interference in Adaptive AUTOSAR-based platforms, with a focus on managing Memory, Timing, and Execution challenges. It explores the robust safety mechanisms in Classic AUTOSAR that ensure Freedom from Interference and the significant challenges in achieving interference-free operation in Adaptive AUTOSAR environments while adhering to ISO26262 standards. The study emphasizes strategies for managing complexities and outlines the multifaceted landscape of achieving interference-free operation. Additionally, it discusses ASIL-compliant Hypervisor, memory partitioning, and Platform Health Management as mechanisms for ensuring safety execution. The paper also raises open questions regarding real-time problems in live projects that are not solved with existing safety mechanisms. Adaptive AUTOSAR plays a crucial role in the development of autonomous and connected vehicles, where functional safety is of utmost
Jain, Yesha
With the growing diversification of modern urban transportation options, such as delivery robots, patrol robots, service robots, E-bikes, and E-scooters, sidewalks have gained newfound importance as critical features of High-Definition (HD) Maps. Since these emerging modes of transportation are designed to operate on sidewalks to ensure public safety, there is an urgent need for efficient and optimal sidewalk routing plans for autonomous driving systems. This paper proposed a sidewalk route planning method using a cost-based A* algorithm and a mini-max-based objective function for optimal routes. The proposed cost-based A* route planning algorithm can generate different routes based on the costs of different terrains (sidewalks and crosswalks), and the objective function can produce an efficient route for different routing scenarios or preferences while considering both travelling distance and safety levels. This paper’s work is meant to fill the gap in efficient route planning for
Bao, ZhibinLang, HaoxiangLin, Xianke
Reproducing driving scenarios involving near-collisions and collisions in a simulator can be useful in the development and testing of autonomous vehicles, as it provides a safe environment to explore detailed vehicular behavior during these critical events. CARLA, an open-source driving simulator, has been widely used for reproducing driving scenarios. CARLA allows for both manual control and traffic manager control (the module that controls vehicles in autopilot manner in the simulation). However, current versions of CARLA are limited to setting the start and destination points for vehicles that are controlled by traffic manager, and are unable to replay precise waypoint paths that are collected from real-world collision and near-collision scenarios, due to the fact that the collision-free pathfinding modules are built into the system. This paper presents an extension to CARLA’s source code, enabling the replay of exact vehicle trajectories, irrespective of safety implications
Ai, YanAdhikari, BikramPark, Chung-KyuKan, Cing-DaoWijesekera, Duminda
Reducing aerodynamic drag through Vehicle-Following is one of the energy reduction methods for connected and automated vehicles with advanced perception systems. This paper presents the results of an investigation aimed at assessing energy reduction in light-duty vehicles through on-road tests of reducing the aerodynamic drag by Vehicle-Following. This study provides insights into the effects of lateral positioning in addition to intervehicle distance and vehicle speed, and the profile of the lead vehicle. A series of tests were conducted to analyze the impact of these factors, conducted under realistic driving conditions. The research encompasses various light-duty vehicle models and configurations, with advanced instrumentation and data collection techniques employed to quantify energy-saving potential. The study featured two sets of L4 capable light duty vehicles, including the Stellantis Pacifica PHEV minivan and Stellantis RAM Truck, examined in various lead and following vehicle
Poovalappil, AmanRobare, AndrewSchexnaydre, LoganSanthosh, PruthwirajBahramgiri, MojtabaBos, Jeremy P.Chen, BoNaber, JeffreyRobinette, Darrell
The trend for the future mobility concepts in the automotive industry is clearly moving towards autonomous driving and IoT applications in general. Today, the first vehicle manufacturers offer semi-autonomous driving up to SAE level 4. The technical capabilities and the legal requirements are under development. The introduction of data- and computation-intensive functions is changing vehicle architectures towards zonal architectures based on high-performance computers (HPC). Availability of data-connection to the backend and the above explained topics have a major impact on how to test and update such ‘software-defined’ vehicles and entire fleets. Vehicle diagnostics will become a key element for onboard test and update operations running on HPCs, as well as for providing vehicle data to the offboard backend infrastructure via Wi-Fi and 5G at the right time. The standard for Service Oriented Vehicle Diagnostics (SOVD) supports this development. It describes a programming interface for
Mayer, JulianBschor, StefanFieth, Oliver
Several challenges remain in deploying Machine Learning (ML) into safety critical applications. We introduce a safe machine learning approach tailored for safety-critical industries including automotive, autonomous vehicles, defense and security, healthcare, pharmaceuticals, manufacturing and industrial robotics, warehouse distribution, and aerospace. Aiming to fill a perceived gap within Artificial Intelligence and ML standards, the described approach integrates ML best practices with the proven Process Failure Mode & Effects Analysis (PFMEA) approach to create a robust ML pipeline. The solution views ML development holistically as a value-add, feedback process rather than the resulting model itself. By applying PFMEA, the approach systematically identifies, prioritizes, and mitigates risks throughout the ML development pipeline. The paper outlines each step of a typical pipeline, highlighting potential failure points and tailoring known best practices to minimize identified risks. As
Schmitt, PaulSeifert, Heinz BodoBijelic, MarioPennar, KrzysztofLopez, JerryHeide, Felix
Artificial Intelligence has gained lot of traction and importance in the 21st century with use cases ranging from speech recognition, learning, planning, problem solving to search engines etc. Artificial Intelligence also has played a key role in the development of autonomous vehicles and robots ranging from perception, localization, decision to controls. Within the big AI umbrella there is machine learning which is all about using your computer to "learn" how to deal with problems without “programming". Deep learning is a branch of machine learning based on a set of algorithms that learn to represent the data directly from the input such as an image, text, Sound, etc. Within deep learning there are Convolutional Neural Networks and Recurrent Neural Networks (CNN/RNN). The study here used convolutional neural network approach to perform image/object recognition. Given that the objective of the autonomous or semi-autonomous vehicle is to promote safety and reduce number of accidents, it
Mansour, IyadSingh, Sanjay
To ensure the safety and stability of road traffic, autonomous vehicles must proactively avoid collisions with traffic participants when driving on public roads. Collision avoidance refers to the process by which autonomous vehicles detect and avoid static and dynamic obstacles on the road, ensuring safe navigation in complex traffic environments. To achieve effective obstacle avoidance, this paper proposes a CL-infoRRT planning algorithm. CL-infoRRT consists of two parts. The first part is the informed RRT algorithm for structured roads, which is used to plan the reference path for obstacle avoidance. The second part is a closed-loop simulation module that incorporates vehicle kinematics to smooth the planned obstacle avoidance reference path, resulting in an executable obstacle avoidance trajectory. To verify the effectiveness of the proposed method, four static obstacle test scenarios and four RRT comparison algorithms were designed. The implementation results show that all five
Wu, WeiLu, JunZeng, DequanYang, JinwenHu, YimingYu, QinWang, Xiaoliang
In this study, we introduce RGB2BEV-Net, an end-to-end pipeline that extends traditional BEV segmentation models by utilizing raw RGB images with Bird’s Eye View (BEV) generation. While previous work primarily focused on pre-segmented images to generate corresponding BEV maps, our approach expands this by collecting RGB images alongside their affiliated segmentation masks and BEV representations. This enables direct input of RGB camera sensors into the pipeline, reflecting real-world autonomous driving scenarios where RGB cameras are commonly used as sensors, rather than relying on pre-segmented images. Our model processes four RGB images through a segmentation layer before converting them into a segmented BEV, implemented in the PyTorch framework after being adapted from an original implementation that utilized a different framework. This adaptation was necessary to improve compatibility and ensure better integration of the entire system within autonomous vehicle applications. We
Hossain, SabirLin, Xianke
Recent years have seen a strong move towards Software Defined Vehicles (SDV) concept as it is seen as an enabler for advancing the mobility by integrating complex technologies like Artificial Intelligence (AI) and Connected Autonomous Driving (CAD) into the vehicle. However, this comes with fundamental changes to the vehicle’s Electrical/Electronic (EE) architecture which require novel testing approaches. This paper presents FEV’s SDV Hardware-In- The-Loop (HIL) test setup which focuses on testing the developed HPC-based software. The functionality of the SDV HIL test setup is demonstrated by testing the software of multiple technologies within the High Performance Computer (HPC) environment like ADAS and teleoperation virtual control units with Over-the-air (OTA) up- dates capability. Test results show the effectiveness of utilizing FEV’s HIL setup in developing and validating the software of SDV platforms.
Obando, DavidAlzu'bi, HamzehCarreón Vásquez, ErwinAlrousan, QusayAlnajdawi, Mohammad SamiTasky, Thomas
Traditional methods for developing and evaluating autonomous driving functions, such as model-in-the-loop (MIL) and hardware-in-the-loop (HIL) simulations, heavily depend on the accuracy of simulated vehicle models and human factors, especially for vulnerable road user safety systems. Continuation of development during public road deployment forces other road users including vulnerable ones to involuntarily participate in the development process, leading to safety risks, inefficiencies, and a decline in public trust. To address these deficiencies, the Vehicle-in-Virtual-Environment (VVE) method was proposed as a safer, more efficient, and cost-effective solution for developing and testing connected and autonomous driving technologies by operating the real vehicle and multiple other actors like vulnerable road users in different test areas while being immersed within the same highly realistic virtual environment. This VVE approach synchronizes real-world vehicle and vulnerable road user
Chen, HaochongCao, XinchengGuvenc, LeventAksun Guvenc, Bilin
Items per page:
1 – 50 of 2807