Search

Distinguishing identity-unrelated background information from discriminative identity information poses a challenge in unsupervised vehicle re-identification (Re-ID). Re-ID models suffer from varying degrees of background interference caused by continuous scene variations. The recently proposed segment anything model (SAM) has demonstrated exceptional performance in zero-shot segmentation tasks. The combination of SAM and vehicle Re-ID models can achieve efficient separation of vehicle identity and background information. This paper proposes a method that combines SAM-driven mask autoencoder (MAE) pre-training and background-aware meta-learning for unsupervised vehicle Re-ID. The method consists of three sub-modules. First, the segmentation capacity of SAM is utilized to separate the vehicle identity region from the background. SAM cannot be robustly employed in exceptional situations, such as those with ambiguity or occlusion. Thus, in vehicle Re-ID downstream tasks, a spatially-constrained vehicle background segmentation method is presented to obtain accurate background segmentation results. Second, SAM-driven MAE pre-training utilizes the aforementioned segmentation results to select patches belonging to the vehicle and to mask other patches, allowing MAE to learn identity-sensitive features in a self-supervised manner. Finally, we present a background-aware meta-learning method to fit varying degrees of background interference in different scenarios by combining different background region ratios. Our experiments demonstrate that the proposed method has state-of-the-art performance in reducing background interference variations. [ABSTRACT FROM AUTHOR]

Creating accurate emission models capable of capturing the variability and dynamics of modern propulsion systems is crucial for future mobility planning. This paper presents a methodology for creating THC and NOx emission models for vehicles equipped with start–stop technology. A key aspect of this endeavor is to find techniques that accurately replicate the engine's stop stages when there are no emissions. To this end, several machine learning techniques were tested using the Python programming language. Random forest and gradient boosting methods demonstrated the best predictive capabilities for THC and NOx emissions, achieving R2 scores of approximately 0.9 for engine emissions. Additionally, recommendations for effective modeling of such emissions from vehicles are presented in the paper. [ABSTRACT FROM AUTHOR]

This paper presents the process of creating a model for electric vehicle (EV) energy consumption, enabling the rapid generation of results and the creation of energy maps. The most robust validation indicators were exhibited by an artificial intelligence method, specifically neural networks. Within this framework, two predictive models for EV energy consumption were developed for winter and summer conditions, based on actual driving cycles. These models hold particular significance for microscale road analyses. The resultant model, for test data in summer conditions, demonstrates validation indicators of an R2 of 86% and an MSE of 1.4, while, for winter conditions, its values are 89% and 2.8, respectively, confirming its high precision. The paper also presents exemplary applications of the developed models, utilizing both real and simulated microscale data. The results obtained and the presented methodology can be especially advantageous for decision makers in the management of city roads and infrastructure planners, aiding both cognitive understanding and the better planning of charging infrastructure networks. [ABSTRACT FROM AUTHOR]

The perception system is essential for autonomous vehicle safety and stability. However, on‐board sensors are vulnerable to external attacks, compromising driving strategies and security. Traditional attack detection methods, relying on static anomaly thresholds from data distributions, falter in complex driving scenarios. Recent studies have sought more nuanced detection techniques, including deep learning and inter‐vehicle communication, but these approaches face limitations related to data security and environmental dependency. This letter introduces an innovative, adaptive method for sensor attack detection in autonomous vehicles. By modelling the perception system's functions and employing the Gaussian process for probabilistic modelling, we generate dynamic uncertainty estimates that serve as adaptable anomaly boundaries. Our method's efficacy and resilience are validated through quantitative analysis of the real‐world KITTI dataset, demonstrating superior adaptive detection capabilities. [ABSTRACT FROM AUTHOR]

Current emission models primarily focus on traditional combustion vehicles and may not accurately represent emissions from the increasingly diverse vehicle fleet. The growing presence of hybrid and electric vehicles requires the development of accurate emission models to measure the emissions and energy consumption of these vehicles. This issue is particularly relevant for low-emission zones within cities, where effective mobility planning relies on simulation models using continuously updated databases. This research presents a two-dimensional emission model for hybrid vehicles, employing artificial neural networks for low-emission zones. The key outcome is the methodology developed to create a CO2 emission model tailored for hybrid vehicles, which can be used to simulate various road solutions. The CO2 emission model achieved an R2 coefficient of 0.73 and an MSE of 0.91, offering valuable information for further advancements in emission modelling. [ABSTRACT FROM AUTHOR]

In order to perform optimizations with vehicle models efficiently, the resource requirements for a simulation must be reduced. In this article, approaches for the physical reduction of the vehicle model are presented, which automatically removes components or elements from the model. For the reduction, evaluation functions considering the crash behaviour of the model are used, as well as geometric methods that ensure its viability. Suitable boundary conditions, such as an equivalent mass or displacements of individual nodes, are applied to the reduced model to recover the kinematics. Finally, the practical application of the methodology is presented by two optimizations. [ABSTRACT FROM AUTHOR]

Nowadays, the topic of reducing vehicles' energy consumption is very important. In particular, for electric vehicles, the reduction of energy consumption is necessary to remedy the most critical problems associated with this type of vehicle: the problem of the limited range of the electric traction, also associated with the long recharging times of the battery packs. To reduce use-phase impacts and energy consumptions of vehicles, it is useful to reduce the vehicle mass (lightweighting). The aim of this work is to analyze the parameters of a vehicle which influence the results of lightweighting, in order to provide guidelines for the creation of a vehicle model suitable for studying the effects of lightweighting. This study was carried out through two borderline case models, a compact car and an N1 vehicle, and simulating these through a consolidated vehicle simulation tool useful for consumption estimations. This study shows that the parameters that most influence the outcome of lightweighting are the rolling resistance, the battery pack characteristics, the aerodynamic coefficients, and the transmission efficiency, while the inertia contributions can be considered negligible. An analysis was also carried out with the variation of the driving cycle considered. [ABSTRACT FROM AUTHOR]

This paper presents a state-of-the-art estimation technique by cross-combining a number n of filters for high-precision, reliable and robust vehicle sideslip angle state estimation, over a full range of vehicle operations irrespective of the driving mission and disruptions that may occur in the system. A machine-learning algorithm based on decision trees connects several filters together to switch between them according to the driving context, ensuring the best possible state estimate for relatively small and large sideslip angle values. In conjunction with the above-mentioned aspects, a seamless transition between different vehicle models is attained by observing the key parameters characterizing the lateral motion of the vehicle. The tests conducted using a prototype vehicle on a snow-covered track confirm the effectiveness and reliability of the proposed approach. [ABSTRACT FROM AUTHOR]

Passive suspensions cannot sufficiently reduce vehicle vibrations on tracks with poor irregularities or when the vehicle was operated at higher speeds. This study proposes three semi-active control algorithms for a secondary lateral damper to enable the vehicle to be operated on high-speed dedicated railways at 350 km/h and conventional low-speed railways at 160 km/h. A high-speed rail vehicle dynamics model with controllable dampers is built to study the feasibility of those algorithms through numerical investigations. Then, both field tests on track and lab tests on a full-scale roller rig were performed to examine the functionality of the controller hardware system and the valve-driven semi-active damper. Furtherly, a time-delay compensation based on the signal amplitude is developed and verified. Results from the on-track and roller rig tests state that the proposed semi-active control algorithms and the designed control system can effectively improve the ride quality of the vehicle when it is put into operation in different track conditions and at various speeds. [ABSTRACT FROM AUTHOR]

In this paper, we present a study of the license plate recognition (LPR) system for Taiwanese vehicles using a cascade of You Only Look Once version 4 (YOLOv4) detectors. The LPR system is composed of a vehicle detection model, a license plate (LP) detection model, an LP corner prediction model, and an LPR model. Herein, the pretrained YOLOv4 model was directly applied to vehicle detection. The YOLOv4 framework was adopted in the LP detection and LP recognition models, performing transfer learning on each model. Furthermore, to enhance the accuracy of the LPR system, an LP corner prediction model was developed in this study to predict the four corner positions of an LP to perform a perspective transformation on the plate for alignment purposes. The experimental results show that our LPR system achieves an accuracy of 98.88% when tested on 2049 images of the application-oriented LP dataset, outperforming most LPR systems reported in the literature. [ABSTRACT FROM AUTHOR]

This study combines laboratory experiments and numerical modelling in a novel manner to assess vehicle stability. Assessing vehicle stability forms the basis of hazard classification criteria, which in turn helps in forming safety guidelines for stream crossings and planning of evacuation routes in times of floods. These criteria are based on theoretical and physical model studies carried out on different vehicle models. This article demonstrates the application of a numerical method to determine the vehicle stability threshold so that the need for a physical model study for each kind of vehicle may be avoided. The numerical investigation is performed using smoothed particle hydrodynamics (SPH) with the vehicle oriented perpendicular to the flow direction, as this is the most critical orientation. A physical model study is also performed and its results are used to validate the SPH model. The results confirm the current Australian Rainfall and Runoff (ARR) safety criteria for stationary vehicles. It also suggests that the ARR stability curve can shift depending on the road conditions that affect the vehicle's sliding mechanism. [ABSTRACT FROM AUTHOR]

To deal with the threat from flying vehicles, it is of great significance to accurately predict the flight trajectory of flying vehicles using the known historical position data. In this paper, we investigate eight typical types of maneuvers of flying vehicles. Then, according to the kinematics model of the flying vehicle, eight kinds of typical maneuver trajectory equations are introduced. Next, the data set for neural network training is created by varying the critical parameters of the trajectory equation. Accordingly, we train the offline Long-Short Term Memory (LSTM) using 1/12 of the dataset of trajectories, and results show that the trained network can classify types of trajectories accurately. Based on the results of classification using the offline LSTM, we propose two trajectory prediction methods, and both of them achieve excellent prediction with and without noise interference. Furthermore, we propose an online prediction method, and it can accurately predict the subsequent trajectory of flying vehicles when the type of trajectory is not clear. And we propose a filter to improve the accuracy of the online prediction method for trajectory prediction with noise. The simulation results verify that all three previously proposed methods can accurately predict the next move based on historical data. [ABSTRACT FROM AUTHOR]

Accurate estimations and assessments of vehicle emissions can support decision-making processes. Current emission estimation tools involve several calculation methods that provide estimates of the exhaust components that result from driving on urban arterial roads. This is an important consideration, as the emissions generated have a direct impact on the health of pedestrians near the roads. In recent years, there has been an increase in the use of emission models, especially in combination with traffic simulator models. This is because it is very difficult to obtain an actual measurement of road emissions for all vehicles travelling along the analysed road section. This paper concerns a review of selected traffic simulations and the estimation of exhaust gas components models. The models presented have been aggregated into a group with respect to their scale of accuracy as micro, meso, and macro. This paper also presents an overview of selected works that combine both traffic and emission models. The presented literature review also emphasises the proper calibration process of simulation models as the most important factor in obtaining accurate estimates. This work also contains information and recommendations on modelling that may be helpful in selecting appropriate emission estimation tools to support decision-making processes for, e.g., road managers. [ABSTRACT FROM AUTHOR]

The vehicle logo contains the vehicle's identity information, so vehicle logo detection (VLD) technology has extremely important significance. Although the VLD field has been studied for many years, the detection task is still difficult due to the small size of the vehicle logo and the background interference problem. To solve these problems, this paper proposes a method of VLD based on the YOLO-T model and the correlation of the vehicle space structure. Aiming at the small size of the vehicle logo, we propose a vehicle logo detection network called YOLO-T. It integrates multiple receptive fields and establishes a multi-scale detection structure suitable for VLD tasks. In addition, we design an effective pre-training strategy to improve the detection accuracy of YOLO-T. Aiming at the background interference, we use the position correlation between the vehicle lights and the vehicle logo to extract the region of interest of the vehicle logo. This measure not only reduces the search area but also weakens the background interference. We have labeled a new vehicle logo dataset named LOGO-17, which contains 17 different categories of vehicle logos. The experimental results show that our proposed method achieves high detection accuracy and outperforms the existing vehicle logo detection methods. [ABSTRACT FROM AUTHOR]

Because the traditional computing model can no longer meet the particularity of Internet of Vehicles tasks, aiming at its characteristics of high bandwidth, low latency and high reliability, this paper proposes a resource allocation strategy for Internet of Vehicles using reinforcement learning in edge cloud computing environment. First, a multi-layer resource allocation model for Internet of Vehicles is proposed, which uses the cooperation mode of edge cloud computing servers and roadside units to dynamically coordinate edge computing and content caching. Then, based on the construction of communication model, calculation model and cache model, make full use of idle resources in Internet of Vehicles to minimize network delay under the condition of limited energy consumption. Finally, the optimization goal is solved by two-layer deep Q network model, and the best resource allocation plan is obtained. The simulation results based on the Internet of Vehicles model show that the computational energy consumption and system delay of proposed strategy do not exceed 400 J and 600 ms, respectively. Besides, the overall effect of resource allocation is better than other comparison strategies and it has certain application prospects. [ABSTRACT FROM AUTHOR]

Vehicle identification and re-identification is an essential tool for traffic surveillance. However, with cameras at every corner of the street, there is a requirement for private surveillance. Automated surveillance can be achieved through computer vision tasks such as segmentation of the vehicle, classification of the make and model of the vehicle and license plate detection. To achieve a unique representation of every vehicle on the road with just the region of interest extracted, instance segmentation is applied. With the frontal part of the vehicle segmented for privacy, the vehicle make is identified along with the license plate. To achieve this, a dataset is annotated with a polygonal bounding box of its frontal region and license plate localization. State-of-the-art methods, maskRCNN, is utilized to identify the best performing model. Further, data augmentation using multiple techniques is evaluated for better generalization of the dataset. The results showed improved classification as well as a high mAP for the dataset when compared to previous approaches on the same dataset. A classification accuracy of 99.2% was obtained and segmentation was achieved with a high mAP of 99.67%. Data augmentation approaches were employed to balance and generalize the dataset of which the mosaic-tiled approach produced higher accuracy. [ABSTRACT FROM AUTHOR]

This paper deals with the control of a hydraulically interconnected suspension system to improve the handling performance of the vehicle. A seven-degrees-of-freedom MATLAB/Simulink vehicle model and an Adams/Car model are used for simulation purposes. A Simscape model for analysing the hydraulic subsystem is also employed. It was observed that the yaw rate and roll responses of the vehicle are influenced to a large extent by the flow control valve opening values. The handling performance of the vehicle was investigated by the control of the orifice valves and their adjustment according to the lateral acceleration inputs. Considerable improvements were obtained in the handling qualities without sacrificing the ride responses, compared to the responses of a passive interconnected suspension. [ABSTRACT FROM AUTHOR]

Next-generation accurate vehicle localization and connectivity technologies will enable significant improvements in vehicle dynamics control. This study proposes a novel control function, referred to as pre-emptive braking, which imposes a braking action if the current vehicle speed is deemed safety-critical with respect to the curvature of the expected path ahead. Differently from the implementations in the literature, the pre-emptive braking input is designed to: a) enhance the safety of the transient vehicle response without compromising the capability of reaching the cornering limit, which is a significant limitation of the algorithms proposed so far; and b) allow — in its most advanced implementation — to precisely constrain the sideslip angle to set levels only through the pre-emptive control of the longitudinal vehicle dynamics, without the application of any direct yaw moment, typical of conventional stability control systems. To this purpose, a real-time-capable nonlinear model predictive control (NMPC) formulation based on a double track vehicle prediction model is presented, and implemented in its implicit form, which is applicable to both human-driven and automated vehicles, and acts as an additional safety function to compensate for human or virtual driver errors in extreme conditions. Its performance is compared with that of: i) two simpler — yet innovative with respect to the state-of-the-art — pre-emptive braking controllers, namely an NMPC implementation based on a dynamic point mass vehicle model, and a pre-emptive rule-based controller; and ii) a benchmarking non-pre-emptive rule-based trail braking controller. The benefits of pre-emptive braking are evaluated through vehicle dynamics simulations with an experimentally validated vehicle model, as well as a proof-of-concept implementation on an automated electric vehicle prototype. [ABSTRACT FROM AUTHOR]

For vehicle state estimation, the conventional Kalman filter performs well under the Gaussian assumption, but in the real non-Gaussian situation (especially when the noise is non-Gaussian heavy-tailed), it shows poor accuracy and robustness. In this paper, an extended Kalman filter (EKF) algorithm based on the maximum correntropy criterion (MCC) is proposed (MCCEKF), and a lateral-longitudinal coupled vehicle model is established, while a state observer containing the yaw rate, vehicle sideslip angle, and longitudinal vehicle speed is designed using the easily available measurement information of on-board sensors. After analyzing the complexity of the proposed algorithm, the new algorithm is verified on the Simulink/CarSim simulation experimental platform by Double Lane Change and Sine Sweep Steer Torque Input maneuver. Experimental results show that the MCC-based EKF algorithm has stronger robustness and better estimation accuracy than the traditional EKF algorithm in the case of non-Gaussian noise, and the MCCEKF is more applicable for vehicle state estimation in practical situations. [ABSTRACT FROM AUTHOR]

Purpose: Metro wheels running on different lines can undergo wear at different positions. This paper aims to investigate the effects of wheel wear at two typical positions, i.e. wheel flange and tread, on the dynamic performance of metro vehicles and analyzes the differences, with an aim of providing theoretical support on wheel reprofiling for different metro lines. Design/methodology/approach: Wheel profile data were measured on two actual metro lines, denoted A and B. It was observed that wheel wear on Lines A and B was concentrated on flanges and treads, respectively. A metro vehicle dynamics model was built using multibody dynamics software SIMPACK. Then it was applied to analyze the differences in effects of wheel wear at different positions on vehicle dynamic performance (VDP) for various speeds (50, 60 and 70 km/h) and line conditions (straight line, R1000m, R600m and R300m curves). Critical speed and vibration acceleration were used as indicators of VDP during linear motion (on straight track), while VDP during curvilinear motion (on curved track) was evaluated in terms of wheel/rail lateral force, wheel/rail vertical force, derailment coefficient and wheel unloading rate. Findings: First, compared to wheel profile with tread wear, wheel profile with flange wear showed better performance during linear motion. When the distance traveled reached 8 × 104 and 14 × 104 km, the vehicle's critical speed was 12.2 and 21.6% higher, respectively. The corresponding vertical and lateral vibration accelerations were 59.7 and 74.8% lower. Second, compared to wheel profile with flange wear, that with tread wear showed better performance during curvilinear motion, with smaller wheel/rail lateral force, derailment coefficient and wheel unloading rate. When the vehicle speed was 50, 60 and 70 km/h, the maximum difference in the three indicators between the two wheel profiles was 40.2, 44.7 and 23.1%, respectively. For R1000m, R600m and R300m curves, the corresponding maximum difference was 45.7, 69.0 and 44.4%, respectively. Practical implications: The results of the study can provide a guidance and theoretical support on wheel reprofiling for different metro lines. On lines with large proportions of curved sections, metro vehicles are more prone to wheel flange wear and have poorer dynamic performance during curvilinear motion. Therefore, more attention should be paid to flange lubrication and maintenance for such lines. On lines with higher proportions of straight sections, metro vehicles are more prone to tread wear and have poorer performance on straight sections. So, tread maintenance and service requires more attention for such lines. Originality/value: Existing research has focused primarily on the effects of wheel wear on VDP, but fails to consider the differences in the effects of wheel wear at different positions on VDP. In actual metro operation, the position of wheel wear can vary significantly between lines. Based on measured positions of wheel wear, this paper examines the differences in the effects of wheel wear at two typical positions, i.e. tread and flange, on VDP in detail. [ABSTRACT FROM AUTHOR]

To explore the mixed traffic characteristics of vehicles and aircraft on the airport surface and solve the problem of real-time conflict detection at key intersections. According to the actual taxiing procedures and airport control rules in China, this paper focus on abstracting the mixed motion process of aircraft and vehicles in the maneuvering area, defining the convergent cross-safety operation scenario. To quantify the driver's attention to safety separation and the degree of conservatism in adjusting speed, the vehicle deceleration rate and acceleration rate are defined with α as the exponent. Under the same spacing, the vehicle deceleration rate is directly proportional to α, and α is named the vehicle safety sensitivity. At the same time, the rules of speed change of vehicles and aircraft will be designed, and a convergent operation model of aircraft and vehicles will be proposed. Based on real-time speed and separation dynamic assessment of safety and efficiency under different traffic strategies, the computer simulation results show that vehicle safety sensitivity and deceleration rules determine the sequence of vehicles and aircraft when they are passing through in the short term and can affect the proportion of mixed traffic flow in the long run. The safety probability of vehicles passing the intersection first is negatively correlated with vehicle safety sensitivity, while the safety probability of aircraft passing the intersection first is correlated positively with vehicle safety sensitivity. The efficiency of passing without conflict with mixed traffic has an inverse relationship with vehicle safety sensitivity. When the vehicle safety sensitivity takes the value in the interval [0.4, 0.6], a mixed traffic flow with higher safety and efficiency, better stability, and a balanced ratio of locomotives can be obtained. [ABSTRACT FROM AUTHOR]

In control theory, a controller parameter-tuning method called virtual reference feedback tuning (VRFT) is proposed. VRFT can tune a feedback controller using single experimental data. If VRFT can be applied to a vehicle, the yaw-rate controller can be adequately tuned without the need for a vehicle model. Therefore, this study expands the results of VRFT to realize the desired yaw-rate control of autonomous vehicles. The key point of the proposed method is to realize the auto-tuning of the gain-schedule proportional-integral (PI) controller based on the vehicle velocity. In this study, the validity of the proposed method using a vehicle dynamics block set is verified. [ABSTRACT FROM AUTHOR]

The longitudinal distance between the vehicle and the forward vehicle, as well as the longitudinal distance between the vehicle and the opposite vehicle, is the main risk factor of overtaking behavior on two-way two-lane roads. Accurate measurement of these distances is the basis and key to automatic driving technology of two-way two-lane roads. In order to measure these longitudinal distances and improve the ranging accuracy, a vehicle distance measurement method of two-way two-lane roads based on monocular vision was proposed. Firstly, the vehicle detection model suitable for two-way two-lane roads was trained using YOLOv5s neural network. Secondly, aiming at the problem that the camera roll angle is not considered in the traditional geometric ranging method, the influence of the roll angle of the camera on ranging results using the traditional geometric ranging method was analyzed. In addition, the improved geometric ranging method considering the roll angle of the camera was proposed. Then, tests were conducted on a two-way two-lane road, and the results showed that the proposed method was effective. Compared with other methods, the improved geometric ranging method has higher ranging accuracy in this scene and can provide a reference for vision-based vehicle distance measurement in multi-lane scenes. [ABSTRACT FROM AUTHOR]

Vehicle re-identification (ReID) is to find a particular vehicle across different surveillant cameras in city roads. There are two main challenges in vehicle ReID. In one case, pictures of the same vehicle from different camera views often look different, which is called "intra-instance difference" in this paper. In another case, different vehicles of the same model and color often look similar from the same camera view, which is called "inter-instance similarity" in this paper. To deal with these challenges, we design a super-resolution-based part collaboration network (SPCN) for vehicle ReID, which aims to improve the performance of ReID system by paying attention to local features with strong differentiation. First, a fine-grained object detection method is used to find the discriminative local parts in vehicle images, and the image qualities of these local parts are enhanced by a super-resolution model with long skip connections. Then, a part collaboration mechanism is designed to realize the weighted fusion of local features, since different local features show different importance in the process of vehicle ReID task. Finally, the global features are fused with the local features to produce the identification results. Extensive experiments on two public data sets show the effectiveness of our method, which introduces less background noise and achieves state-of-the-art results. [ABSTRACT FROM AUTHOR]

In order to solve the problem that the measured values of key state parameters such as the lateral velocity and yaw rate of the vehicle are easily interfered by random errors, a filter estimation method of vehicle state is proposed based on the principle of robust filtering and the unscented particle filter algorithm. Based on the establishment of a 3-DOF non-linear dynamic model and the Dugoff tire model of the vehicle, the adaptive robust unscented particle filter(ARUPF) is used to filter and estimate the parameters of the vehicle state, and to realize the longitudinal and lateral speed as well as the yaw rate of the vehicle during the driving process. The simulation and the real vehicle test results show that based on the adaptive robust unscented particle filter algorithm, the vehicle driving state estimation can be realized, the measurement parameters can be effectively filtered, and the estimation accuracy is high. [ABSTRACT FROM AUTHOR]

We investigate the dynamics of a two‐wheel vehicle model after deviating from an unstable steady powerslide motion. For a model with brush tyre characteristics we observed that for a range of constant driver inputs (front steering angle, driving torque) the motion converges to a periodic oscillation around a circular path with a very small radius of curvature. In this paper we investigate, how a modified tyre model with decreasing friction coefficient for large slip velocities influences the stability and domain of existence for these periodic motions. We also compare the obtained solutions with those using Pacejka's Magic Formula for the tyre characteristics and find a remarkable good agreement. [ABSTRACT FROM AUTHOR]

Accurate acquisition of vehicle dynamics state information is essential for vehicle active safety control systems. However, these states cannot be easily measured, and the measurement is expensive. Conventional Kalman filters perform well for vehicle state estimation in Gaussian environments but exhibit low accuracy and robustness under practical non-Gaussian noise. Vehicle model parameter ingestion, inaccurate tire force calculation, and non-Gaussian noise from on-board sensors cause great challenges to the estimation of vehicle driving states. Therefore, this paper presents a robust hierarchical estimation scheme for vehicle driving state based on the maximum correntropy square-root cubature Kalman filter (MCSCKF) using easily measurable on-board sensor information. First, the vehicle mass is dynamically updated based on the recursive least squares (FRLS) method with a forgetting factor. Then, an adaptive sliding mode observer (ASMO) is designed to estimate the longitudinal and lateral tire forces. Ultimately, the vehicle states are estimated based on the MCSCKF under non-Gaussian noise. Two typical operating situations are carried out to verify the validity of the proposed estimation scheme. The results prove that the proposed estimation scheme can estimate the vehicle's driving state accurately compared to other common methods. And the MCSCKF algorithm has better accuracy and robustness than the traditional Kalman filters for vehicle state estimation in non-Gaussian situations. [ABSTRACT FROM AUTHOR]

To investigate and compare the lane changing behavior of passenger cars and heavy vehicles during the implementation period (defined as the interval from the start time to the end time of a lane change maneuver), this study applies the gradient boosting decision tree (GBDT) method to model the lane changing behavior of heavy vehicles and passenger cars, respectively. Results show that the lane change models vary with the vehicle types and lane change directions. Different factors are considered by the drivers of passenger cars and heavy vehicles when implementing lane changes to different directions. Partial dependence plots of GBDT models reveal that the influence of independent variables on lane changing behavior is nonlinear and complicated, which means that the same variable leads to various effects on the lane change decision across different vehicle types and lane change directions. In contrast with other state-of-the-art methods, the proposed method can obtain more accurate results. The findings indicate that it is necessary to build specific lane change models based on vehicle types and lane change directions for microscopic traffic simulators and autonomous vehicles. [ABSTRACT FROM AUTHOR]

In a half-through or through arch bridge, the hangers (or suspenders) are considered as critical structural components, especially for dynamic performance assessment. This paper aims to study the dynamic impact of stiff hangers in an arch bridge caused by moving vehicles. An iterative algorithm is firstly proposed to solve the three-dimensional coupling vibration between the vehicle(s) and the large-scale bridge. In addition, the uncertainty of the computed strain-based dynamic amplification factor (DAF) due to measurement noise and the relationship between the DAF and local stiffness degeneration are investigated, respectively. After that, the influences of several parameters on the strain-based DAF of the hangers are studied by parametric analyses, including vehicle speed, road surface conditions, loading lane, vehicle spacing and vehicle model. The results show that the proposed iterative algorithm can accurately solve the refined three-dimensional coupled vibration system, and that the strain-based DAF is sensitive to measurement noise and changes of local stiffness. The shortest hanger suffers a more serious impact effect and a higher dynamic fatigue risk. The analysis results of this study could provide some useful and practical recommendations for the routine maintenance and traffic management of arch bridges with stiff hangers. [ABSTRACT FROM AUTHOR]

The inerter is a two-terminal component that can be added to the spring-and-damper configuration of a suspension system. It has the property that the force exerted is proportional to the relative acceleration at its terminals. Studies have demonstrated the inerter's benefit of providing superior vibration isolation when it is used in the vehicle suspension of passenger cars. However, similar benefit on another common vehicle class on the roads, namely heavy vehicles, remain to be shown, as these vehicles have vastly different parameter values than passenger cars. This study is an investigation on the performance improvement brought by an inerter in the suspension of common heavy vehicles. In the study, the parameter values of a truck and a bus were adopted in the quarter vehicle model with two different spring-damper-inerter configurations (parallel and serial inerter), and the improvements in vibration isolation and road holding capability were determined by optimization of inertance. Results show that the inerter is similarly effective in providing the said improvements when implemented on heavy vehicles instead of on passenger cars, judging from reductions in sprung mass acceleration and dynamic tire load. It is also observed that the performance benefit is associated with larger optimum inertance than that for passenger cars. Overall, the inerter has been shown to be beneficial in the parallel and serial configurations, both of which are common and can be practically implemented in the suspension of heavy vehicles. [ABSTRACT FROM AUTHOR]

The Harris Hawk Optimization (HHO) is an excellent algorithm, but it also has the possibility of being trapped in local optima and premature convergence. To increase the accuracy of vehicle parameter identification, an improved HHO is proposed to enhance exploration and exploitation, named MPHHO. The algorithm integrates two unique operations of the Marine Predators Algorithm, which are the strategy affected by fish aggregating devices (FADs) and the marine memory storage strategy. According to the characteristic of HHO that the exploration ability decreases linearly with the advance of the iteration and completely converts to exploitation ability in the second half iteration, the step size control parameter of the strategy considering FADs effect is modified by a concave-convex cos function to further intensify the performance. The superiority of the MPHHO is proved by comparing the results of 23 classic test functions with other advanced nine algorithms. Moreover, the variant of HHO is combined with the transfer matrix method for multibody systems (MSTMM) for the first time to identify the physical parameters of a vehicle model. Good agreements are achieved that the average relative errors of the determined vehicle parameters and the modal parameters are all within 1%, which demonstrates the hybrid method based on the MPHHO and the MSTMM has an excellent performance in vehicle parameter identification. [ABSTRACT FROM AUTHOR]

This paper presents a novel dynamic vehicle tracking framework, achieving accurate pose estimation and tracking in urban environments. For vehicle tracking with laser scanners, pose estimation extracts geometric information of the target from a point cloud clustering unit, which plays an essential role in tracking tasks. However, the point cloud acquired from laser scanners only provides distance measurements to the object surface facing the sensor, leading to nonnegligible pose estimation errors. To address this issue, we take the motion information of targets as feedback to assist vehicle detection and pose estimation. In addition, the heading normalization vehicle model and a robust target size estimation method are introduced to deduce the pose of a vehicle with 2D matched filtering. Furthermore, considering the mobility of vehicles, we utilize the interactive multitude model (IMM) to capture multiple motion patterns. Compared to existing methods in the literature, our method can be applied to spatially sparse or incomplete point cloud observations. Experimental results demonstrate that our vehicle tracking framework achieves promising performance, and its real-time capability is also validated in real traffic scenarios. [ABSTRACT FROM AUTHOR]

This study proposes an approach to the yaw and lateral stability of the vehicle that considers their steady-state trend. This study designs a new linear time-varying model predictive control (LTV-MPC) method based on the 2DOF vehicle model. The tire model curve of the magic formula is simplified into two straight lines to replace the nonlinear characteristic of tires and reduce the solution time of the controller under the premise of a good tracking effect. The lateral stiffness of tires is selected according to the predicted trend of stable state of the vehicle. Simulation results show that, under extreme conditions, the proposed method can effectively track the reference vehicle sideslip angle and yaw rate. The real-time performance of the controller is substantially improved compared with NMPC. [ABSTRACT FROM AUTHOR]

This paper constructs a nonlinearly parametrized hypersonic cruise vehicle model on the basis of the existing on-orbit flight data using a curve-fitting technique. The hypersonic cruise vehicle system is separated into two interconnected subsystem: an attitude subsystem with rotational dynamics and a velocity subsystem with engine dynamics. The continuous adaptive controllers are designed for two subsystems using a novel function bounding technique and appropriate coordinate transformations, respectively, which ensure the global boundedness of all signals and achieve the non-zero equilibrium point regulation of nonlinearly parametrized hypersonic vehicle systems. One of the implications of this result is that growing nonlinearities in the uncertain model of the hypersonic vehicle system may be allowed for global stabilization. A simulation result verifies the effectiveness of the proposed adaptive control scheme. [ABSTRACT FROM AUTHOR]

This paper presents an integrated controller for an autonomous articulated electric vehicle (AAEV) for path tracking and rollover prevention. The AAEV is vulnerable to rollover due to the characteristics of the articulated frame steering (AFS) mechanism, which shows improved maneuverability and agility but not front wheel steering. In addition, the ratio between height and track width is high, so the AAEV is prone to rolling over. Therefore, the proposed controller was designed to achieve the two goals, following the reference path and managing the velocity to improve the safety of the AAEV. Vehicle behavior was modeled by a kinematic model with actuation delay. A local linearization was used to improve the accuracy of the vehicle model and reduce the computational load. Reference states of the position and heading were determined to follow the reference path and prevent the rollover. A model predictive control (MPC)-based reference state tracker was designed to optimize the articulation angle rate and longitudinal acceleration commands. The simulation study was conducted to evaluate the proposed algorithm with a comparison of the base algorithms. The reference path for the simulation was an S-shaped path with discontinuous curvature. Simulation results showed that the proposed algorithm reduces the path tracking error and load-transfer ratio. [ABSTRACT FROM AUTHOR]

This paper proposes the dissipation of the energy analysis approach for vehicle plane motion stability analysis. The vehicle uses a three-degree-of-freedom vehicle model, and a unified tire-sideslip-angle model suitable for large tire-sideslip-angle and wheel reverse rotation conditions. Two calculation methods of total dissipation of energy are proposed: the integral calculation method and the energy difference calculation method. This paper analyses the system stability of the vehicle model, using the total dissipation of the energy indicator and compares the results with the Lyapunov exponent indicator. It shows that the dissipation of the energy analysis approach has a fast calculation speed, strong interpretability and high calculation accuracy, which can intuitively and quantitatively reveal the global stability characteristics of vehicle plane motion. [ABSTRACT FROM AUTHOR]

The high-speed steering control of tracked vehicles has been studied thoroughly, but the high-speed steering stability boundary of tracked vehicles has rarely been investigated. The establishment of a dynamics model with a high computational accuracy is the premise of studying the high-speed steering stability boundary of tracked vehicles. In this work, the track forces of tracked vehicles during steering were determined based on a shear stress model, and the steering dynamics model of tracked vehicles was established. A zero-differential-steering tracked vehicle was used as the object of study. The calculation results of the steering kinematics model and dynamics model of tracked vehicles under different steering conditions were studied in detail on sandy road surface environment, as well as the change laws of the steering trajectory, steering radius, vehicle centroid velocity, and steering slip angle. The steering trajectories of tracked vehicles on five kinds of roads under the same steering conditions were determined by simulations. By setting the critical condition for unstable steering, the corresponding relationship between the maximum circumferential velocity difference of the sprocket Δ u max , the minimum steering radius R min , the maximum sideslip angle β max , and the theoretical centroid speed v t h was obtained when the vehicle turned on five kinds of roads, and it was used to determine the high-speed steering stability boundary of tracked vehicles. The simulation results can provide a reference for the design of tracked vehicle steering mechanisms and high-speed steering control. [ABSTRACT FROM AUTHOR]

Aimed at the safety and stability problems of intelligent vehicles under extreme conditions such as low adhesion road surface and emergency lane change and obstacle avoidance, this article designs a lane change and obstacle avoidance controller based on road adhesion coefficient. Using the nonlinear vehicle dynamics model as the prediction model, using the recursive least squares method to identify the road adhesion coefficient, considering the road adhesion coefficient to plan and adjust in the obstacle avoidance path as well as limit constraint conditions of the model predictive control controller, using model predictive control method for the expectation of intelligent vehicle trajectory tracking, travels tremendously guarantee the security and stability of driving. The joint CarSim–Simulink simulations results show that under poor road conditions, the trajectory tracking accuracy after optimization is higher and the vehicle is less prone to sideslip and instability. The lane change controller designed in this article has strong adaptability to different road surface adhesion coefficient, and all parameters can be controlled within a reasonable safety range at different speeds, with good robustness. [ABSTRACT FROM AUTHOR]

This paper investigates the attitude control problem of the morphing vehicle subject to great dynamics changes and disturbances during the morphing phase. An improved model‐free adaptive control (MFAC) method is proposed based on the compact format dynamic linearization technique. Firstly, a compact format data model and data‐driven control scheme are established based on the input/output (I/O) data of the controlled plant, independent of the complicated and time‐varying mathematical model of the morphing vehicle. Secondly, a series of historical output data errors in a moving time window are introduced to the control law, in which the length of the moving time window can be adjusted according to the system order of controlled plant. The convergence and stability of the improved control law are then proved theoretically, which guarantees the convergence of the tracking errors and the boundedness of the input and output data. Finally, numerical simulations are presented to evaluate the proposed approach, and comparisons are made with the conventional proportional‐derivative control method. Simulation results demonstrate that the improved method possesses better effectiveness and robustness in the presence of model changes and disturbances. [ABSTRACT FROM AUTHOR]

Improving the stability and safety is of great significance for the in-wheel electric vehicle. There are many studies only concentrating on active front steering (AFS) control or direct yaw-moment control (DYC). However, When the in-wheel electric vehicle is under extreme conditions, AFS or DYC alone is not effective. In this paper, an integrated controller of AFS and DYC is proposed. Firstly, the ideal values of yaw rate and sideslip angle can be calculated based on the two-degree-of-freedom vehicle model. Secondly, the AFS controller is obtained according to the backstepping-based fast terminal sliding mode (FTSM). Then, the DYC controller which consists of the upper controller and the lower controller is constructed. The upper controller is developed via the integral-based second-order sliding mode (SOSM). The appropriate torque is outputted to each wheel by the lower controller. Finally, the simulation results show that the actual yaw rate and sideslip angle can approach the ideal ones as closely as possible under the proposed integrated controller. [ABSTRACT FROM AUTHOR]

Intrusion detection is an important defensive measure for automotive communications security. Accurate frame detection models assist vehicles to avoid malicious attacks. Uncertainty and diversity regarding attack methods make this task challenging. However, the existing works have the limitation of only considering local features or the weak feature mapping of multifeatures. To address these limitations, we present a novel model for automotive intrusion detection by spatial–temporal correlation (STC) features of in‐vehicle communication traffic (intrusion detection system [IDS]). Specifically, the proposed model exploits an encoding‐detection architecture. In the encoder part, spatial and temporal relations are encoded simultaneously. To strengthen the relationship between features, the attention‐based convolutional network still captures spatial and channel features to increase the receptive field, while attention‐long short‐term memory builds meaningful relationships from previous time series or crucial bytes. The encoded information is then passed to detector for generating forceful spatial–temporal attention features and enabling anomaly classification. In particular, single‐frame and multiframe models are constructed to present different advantages, respectively. Under automatic hyperparameter selection based on Bayesian optimization, the model is trained to attain the best performance. Extensive empirical studies based on a real‐world vehicle attack data set demonstrate that STC‐IDS has outperformed baseline methods and obtains fewer false‐alarm rates while maintaining efficiency. [ABSTRACT FROM AUTHOR]

Vehicle Re-Identification (ReID) is of great significance for public security and intelligent transportation. Large and comprehensive datasets are crucial for the development of vehicle ReID in model training and evaluation. However, existing datasets in this field have limitations in many aspects, including the constrained capture conditions, limited variation of vehicle appearances, and small scale of training and test set, etc. Hence, a new, large, and challenging benchmark for vehicle ReID is urgently needed. In this paper, we propose a large vehicle ReID dataset, called VERI-Wild 2.0, containing 825,042 images. It is captured using a city-scale surveillance camera system, consisting of 274 cameras covering a very large area over 200 $km^2$ k m 2 . Specifically, the samples in our dataset present very rich appearance diversities thanks to the long time span collecting settings, unconstrained capturing viewpoints, various illumination conditions, diversified background environments, and different weather conditions. Furthermore, to facilitate more practical benchmarking, we define a challenging and large test set containing about 400K vehicle images that do not have any camera overlap with the training set. VERI-Wild 2.0 is expected to be able to facilitate the design, adaptation, development, and evaluation of different types of learning models for vehicle ReID. Besides, we also design a new method for vehicle ReID. We observe that orientation is a crucial factor for feature matching in vehicle ReID. To match vehicle pairs captured from similar orientations, the learned features are expected to capture specific detailed differential information for discriminating the visually similar yet different vehicles. In contrast, features are desired to capture the orientation invariant common information when matching samples captured from different orientations. Thus a novel disentangled feature learning network (DFNet) is proposed. It explicitly considers the orientation information for vehicle ReID, and concurrently learns the orientation specific and orientation common features that thus can be adaptively exploited via an adaptive matching scheme when dealing with matching pairs from similar or different orientations. The comprehensive experimental results show the effectiveness of our proposed method. [ABSTRACT FROM AUTHOR]

Commercial vehicles are more likely to park close to their destinations than passenger vehicles even though sometimes parking violations are inevitable for their freight or service activities. Although clear distinctions between commercial and passenger vehicles lead to different illegal parking behaviors, little is known about the difference in double parking behaviors between the two types. To better utilize limited curb spaces, a better understanding of differences and similarities is required so that corresponding policies can be established accordingly. Therefore, this research aims to quantify the difference in double parking behavior between commercial and passenger vehicles using count models from two different perspectives: (a) an individual vehicle level count model to investigate vehicle attributes influencing the number of double parking violations per vehicle and (b) a ZIP Code level count model to investigate economic and built environment factors affecting double parking behaviors. The results showed that double parking is a 'one-time event' for a passenger vehicle and a 'recurring event' for a commercial vehicle. With respect to economic factors, different types of establishments had different levels of impact on double parking for commercial and passenger vehicles in an area. This research discusses how parking policies and strategies can be established based on the similarities and differences between commercial and passenger vehicles' parking violation behavior. [ABSTRACT FROM AUTHOR]

Aiming at a design for buffering and load reduction configuration for a large-scale (diameter greater than 500 mm) vehicle entering water at high speed (greater than 100 m/s), a numerical model for a vehicle entering water at high speed was employed based on an arbitrary Lagrange-Euler (ALE) algorithm. Combined with modal analysis and shock response spectrum, the influence of the head cap on the dynamic characteristics of the structure was analyzed. The results showed that the peak value and pulse width of the impact load on the vehicle increased with the increase in the speed of water entry. The existence of the head cap increased the complexity of the forces on the vehicle during the process of water entry. The initial formation of the cavity was greatly affected by the head cap. The head cap and the vehicle separated in the later stage of the water entry. During the process of water entry, the shell of the vehicle was mainly compressed and bent and the head cap reduced the deformation. The relevant conclusions of this paper can provide some input for the design of a new buffering structure and vehicle shell. [ABSTRACT FROM AUTHOR]

This paper proposes an autonomous vehicle trajectory tracking system that fully considers road friction. When an intelligent vehicle drives at high speed on roads with different friction coefficients, the difficulty of its trajectory tracking control lies in the fast and accurate identification of road friction coefficients. Therefore, an improved strategy is designed based on traditional recursive least squares (RLS), which is utilized for accurate identification of the friction coefficient. First, the tire force and slip rate required for the estimation of the road friction coefficient by constructing the vehicle dynamics model and tire effective model are calculated. In this paper, a variable forgetting factor recursive least squares (VFF-RLS) method is proposed for the construction of the friction coefficient estimator. Second, the identified results are output to the model predictive controller (MPC) constructed in this paper as a way to improve tire slip angle constraints, to realize the trajectory tracking of the intelligent vehicle. Finally, the joint simulation test results of Carsim and Matlab/Simulink show that the trajectory tracking system based on the VFF-RLS friction coefficient estimator has outstanding tracking performance. [ABSTRACT FROM AUTHOR]

In order to improve the safety and reliability of the hypersonic flight vehicle, a sliding mode observer-based fault detection scheme is applied in this paper to handle the actuator fault detection issue, including stuck fault detection and PLOE fault detection. A dynamic linear model with uncertainty is first derived from the original nonlinear hypersonic flight vehicle model by using Taylor's linearization approach at the equilibrium point. Secondly, the actuator fault model, reflecting stuck faults and PLOE faults, is constructed. Then, a sliding mode-based fault detection observer, considering system decomposition, is developed based on the linearized hypersonic flight vehicle model. At last, the designed sliding mode observer is applied to the original nonlinear hypersonic flight vehicle for single-input, single-style actuator fault detection. The simulation results show that stuck faults and big proportion PLOE faults can be timely and accurately detected at the fault time, and the stuck actuator fault from input 3 can cause a deadly impact to the hypersonic flight vehicle, which deserves much more attention than the actuator faults from the other three inputs. Meanwhile, the detection of a small proportion of PLOE faults encounters some difficulties and needs special attention and further investigation. [ABSTRACT FROM AUTHOR]

The present paper is concerned with the wayset-based guidance of underactuated multirotor aerial vehicles (MAVs). A hierarchical guidance and control structure is first established, in which the guidance is realized as a supervisory loop. The lower-level stabilizing attitude and position control laws are assumed to be available. On the other hand, the outer-loop guidance is designed based on a fixed-horizon tube-based robust model predictive control (MPC), which conducts the MAV to visit a given sequence of waysets, without violating their state and control bounds, and allowing the vehicle to rest in each wayset for a specified period. The MPC is designed using a reduced-order closed-loop dynamic model describing the vehicle's translation, which is derived considering the stabilizing position and attitude control laws and the assumption of a time-scale separation between the closed-loop translational and rotational dynamics. This model is put into a discrete-time linear state–space representation subject to additive bounded random disturbance and measurement noise. The properties of the proposed method, which includes the MPC recursive feasibility and robust stability as well as the overall guidance feasibility, are analytically studied. The method is also numerically evaluated using a realistic quadrotor dynamic model, showing its effectiveness and confirming its properties. • Wayset-based supervisory guidance loop for underactuated multirotor aerial vehicles. • Wayset transitions based on dual-mode tube-based robust model predictive control. • Proof of recursive feasibility, robust stability, and overall guidance resolvability. • Obtained performance and low computational burden encourage real-time implementation. [ABSTRACT FROM AUTHOR]

This paper investigates the trajectory tracking control problem for underactuated underwater vehicles, for which a model is expressed in terms of quasi-velocities arising from the inertia matrix decomposition. The control approach takes into account non-modeled dynamics and external disturbances and is suitable for symmetric vehicles. It is shown that such systems can be diagonalized using inertial quasi-velocities (IQVs). The strategy consists of the velocity controller and two adaptive integral sliding mode control algorithms. The proposed approach, introducing velocity transformation and using backstepping methods and integral sliding mode control, allows trajectory tracking for vehicles in described models with symmetric inertia matrix. Proof of the stability of the closed system was carried out using IQV. The proposed scheme has been verified on two 3 DOF models of underwater vehicles with thruster limitations. A brief discussion of the results is also given. [ABSTRACT FROM AUTHOR]

Directional performance and highway stability are two important aspects that need to be considered in development and design of a heavy articulated vehicles. To improve the maneuverability and stability of a multi-trailer articulated heavy vehicle (MTAHV), an active linear quadratic regulator (LQR) multi-axle-steering method is designed and examined. First, a linear yaw-plane model with four-degree-of freedom (4-DOF) for MTAHV is built and validated. Thus, a reference model supplying the desired state responses is introduced. Then, an active control algorithm of multi-axle-steering for the rear axles of tractor and full-trailer is investigated, and a LQR controller is proposed based on the linear vehicle model to make the control variables track the desired state responses. The control strategy concentrates on keeping the actual yaw rate and side-slip angle follow the steady-state yaw rate and zero side-slip angle. Finally, the effectiveness of the designed approach on enhancing the maneuverability and stability of the MTAHV have been validated through the simulations of the low-speed 360° roundabout and the single lane-change maneuver with high speed, respectively. The method has a certain reference value for improving the active safety of the MTAHV. [ABSTRACT FROM AUTHOR]

With the rise of unmanned driving and intelligent transportation research, great progress has been made in vehicle detection technology. The purpose of this paper is employing the method of deep learning to study the vehicle detection algorithm, in which primary-stage target detection algorithms, namely, YOLOv3 algorithm and SSD algorithm, are adopted. Therefore, the method first processes the image data in the open-source road vehicle dataset for training. Then, the vehicle detection model is trained by using YOLOv3 and SSD algorithms to show the detection effect, respectively. The result is by comparing the detection effects of the two models on vehicles. The researchers accomplished the result analysis and summarized the characteristics of various models obtained by training, to apply to target tracking, semantic segmentation, and unmanned driving. [ABSTRACT FROM AUTHOR]