Browse Topic: Education and training

Items (6,264)
Image dehazing techniques can play a vital role in object detection, surveillance, and accident prevention, especially in scenarios where visibility is compromised because of light scattering by atmospheric particles. To obtain a high-quality image or as an initial step in processing, it’s crucial to restore the scene’s information from a single image, given that this is an ill-posed inverse problem. The present approach utilized an unsupervised learning approach to predict the transmission map from a hazy image and used YOLOv8n to detect the car from a clear recovered image. The dehazing model utilized a lightweight parallel channel architecture to extract features from the input image and estimate the transmission map. The clear image is recovered using an atmospheric scattering model and given to the YOLOv8n for car detection. By incorporating dark channel prior loss during training, the model eliminates the need for a paired dataset. The proposed dehazing model with fewer
Dave, ChintanPatel, HetalKumar, Ahlad
Bearings are fundamental components in automotive systems, ensuring smooth operation, efficiency, and longevity. They are widely used in various automotive systems such as wheel hubs, transmissions, engines, steering systems etc. Early detection of bearing defects during End-of-Line (EOL) testing and operational phases is crucial for preventive maintenance, thereby preventing system malfunctions. In the era of Industry 4.0, vibrational, accelerometer, and other IoT sensors are actively engaged in capturing performance data and identifying defects. These sensors generate vast amounts of data, enabling the development of advanced data-driven applications and leveraging deep learning models. While deep learning approaches have shown promising results in bearing fault diagnosis, they often require extensive data, complex model architectures, and specialized hardware. This study proposes a novel method leveraging the capabilities of Vision Language Models (VLMs) and Large Language Models
Chandrasekaran, BalajiCury, Rudoniel
To optimize vehicle chassis handling stability and ride safety, a layered joint control algorithm based on phase plane stability domain is proposed to promote chassis performance under complicated driving conditions. First, combining two degrees-of-freedom vehicle dynamics model considering tire nonlinearity with phase plane theory, a yaw rate and side slip angle phase plane stability domain boundary is drew in real time. Then based on the real-time stability domain and hierarchical control theory, an integrated control system with active front steering (AFS) and direct yaw moment control (DYC) is designed, and the stability of the controller is validated by Lyapunov theory. Finally, the lateral stability of the vehicle is validated by Simulink and CarSim simulations, real car data, and driving simulators under moose test and pylon course slalom test. The experimental results confirm that the algorithm can enhance the maneuverability and ride safety for intelligent vehicles.
Liao, YinshengZhang, ZhijieSu, AilinZhao, BinggenWang, Zhenfeng
For mature virtual development, enlarging coverage of performances and driving conditions comparable with physical prototype is important. The subjective evaluation on various driving conditions to find abnormal or nonlinear phenomena as well as objective evaluation becomes indispensable even in virtual development stage. From the previous research, the road noise had been successfully predicted and replayed from the synthesis of system models. In this study, model based NVH simulator dedicated to virtual development have been implemented. At first, in addition to road noise, motor noise was predicted from experimental models such as blocked force and transfer function of motor, mount and body according to various vehicle conditions such as speed and torque. Next, to convert driver’s inputs such as acceleration and brake pedal, mode selection button and steering wheel to vehicle’s driving conditions, 1-D performance model was generated and calibrated. Finally, the audio and visual
Park, SangyoungDirickx, TomKang, Yeon JuneNam, Jeong MinGonçalves, Vinícius Valencia
There are some paradoxical keys to NVH engineering success that are not taught in engineering schools. This paper will describe these in detail and provide examples to add context. The first unexpected key is that a good generalist makes a better expert. The more you understand the complete product development process, and the better contacts you have throughout the product development organization, the easier it will be for you to find cost effective solutions to your specific issues. Next, you need to know your customers, and that means both internal and external customers. If you work for a supplier, it means knowing original equipment manufacturer (OEM) and end user customers. The more you understand the customers’ needs, the better you can address them and make your product stand out. Another key is to try to turn a crazy idea into something practical. Sometimes you might find a completely insane solution to your problem, such as making a major component out of solid gold. If you
Reinhart, Thomas
A test and signal processing strategy was developed to allow a tire manufacturer to predict vehicle-level interior response based on component-level testing of a single tire. The approach leveraged time-domain Source-Path-Contribution (SPC) techniques to build an experimental model of an existing single tire tested on a dynamometer and substitute into a simulator vehicle to predict vehicle-level performance. The component-level single tire was characterized by its acoustic source strength and structural forces estimated by means of virtual point transformation and a matrix inversion approach. These source strengths and forces were then inserted into a simulator vehicle model to predict the acoustic signature, in time-domain, at the passenger’s ears. This approach was validated by comparing the vehicle-level prediction to vehicle-level measured response. The experimental model building procedure can then be adopted as a standard procedure to aid in vehicle development programs.
Nashio, HiroshiKajiwara, KoheiRinaldi, GiovanniSakamoto, Yumiko
The implementation of active sound design models in vehicles requires precise tuning of synthetic sounds to harmonize with existing interior noise, driving conditions, and driver preferences. This tuning process is often time-consuming and intricate, especially facing various driving styles and preferences of target customers. Incorporating user feedback into the tuning process of Electric Vehicle Sound Enhancement (EVSE) offers a solution. A user-focused empirical test drive approach can be assessed, providing a comprehensive understanding of the EVSE characteristics and highlighting areas for improvement. Although effective, the process includes many manual tasks, such as transcribing driver comments, classifying feedback, and identifying clusters. By integrating driving simulator technology to the test drive assessment method and employing machine learning algorithms for evaluation, the EVSE workflow can be more seamlessly integrated. But do the simulated test drive results
Hank, StefanKamp, FabianGomes Lobato, Thiago Henrique
The process of producing aircraft parts involves the drilling of aluminum alloys. This creates a large amount of chips, which are removed using air, but sometimes they still remain within the holes. This is checked by inspectors through visual inspection. However, the quality of human inspection varies based on skill level and fatigue. Thus, image-based inspection should be used to stabilize and further improve inspection quality. This study aims to build a framework for chip detection based on image processing. Taking into account on-site implementation, the system must have low installation and running costs and be standalone. Therefore, we adopt the KIZKI algorithm, which satisfies these conditions. KIZKI means awareness in Japanese. This is a model of human peripheral vision and saccades. It does not require training like AI and can achieve high-speed and high-performance detection using a low-performance computer. In other words, there is no need for a computer with an expensive
Iinuma, MarinSato, JunyaTsuji, Masahiko
In February, the Joint Interagency Field Experimentation (JIFX) team at the Naval Postgraduate School (NPS) executed another highly collaborative week of rapid prototyping and defense demonstrations with dozens of emerging technology companies. Conducted alongside NPS’ operationally experienced warfighter-students, the event is a win-win providing insight to accelerate potential dual-use applications.
With the advancement of control technology in the automotive field, there is a possibility of cross-system redundant control between various actuators. As for the braking system, current brake-by-wire system often uses mechanical backup braking methods to give the vehicle a certain braking capacity after failure. However, in the mechanical backup braking mode, the brake master cylinder is connected to the supporting wheel cylinder, and the brake assist is lost, which leads to an increase in brake pressure and makes it difficult for the driver to step on the brake pedal. Meanwhile, due to the limitation of the brake master cylinder stroke, the maximum braking deceleration of the vehicle is only 3 m/s2 after the driver fully presses the brake pedal. The above two defects greatly affect the safety of the vehicle during backup braking. To solve the above problems, this article takes electric vehicles as the research object, designs a new type of hydraulic circuit for the braking system
Tian, BoshiLi, LiangLiao, YinshengLv, HaijunHu, ZhimingSun, YueQu, Wenying
In this study, an initial approach using deep reinforcement learning to replicate the complex behaviors of motorcycle riders was presented. Three learning examples were demonstrated: following a target velocity, maintaining stability at low speeds, and following a target trajectory. These examples serve as a starting point for further research. Additionally, the proficiency of the constructed models was examined using rider proficiency evaluation methods developed in previous studies. Initial results indicated that the models have the potential to mimic real rider behaviors; however, challenges such as differences between the model’s output and what humans can produce were also identified for future work.
Mitsuhashi, YasuhiroMomiyama, YoshitakaYabe, Noboru
The authors will present findings from their cradle-to-cradle Product Carbon Footprint (PCF) study which captures an objective and comprehensive system level evaluation of the greenhouse gas (GHG) footprint of four different material types used in the same automotive application: Unsaturated Polyester Resin (UPR) SMC, steel, aluminum and glass fiber reinforced polypropylene (PP-GF). This study includes the simulation driven design of four mid-sized pickup boxes which were designed according to automotive requirements and relevant design guidelines for each material. OEM experts were consulted to validate the relevant specifications and boundary conditions. The technical paper includes details on the geometric design, simulation, production processes, life cycle and environmental impact assessment all in compliance with ISO standards (14040/14044) for the Cradle-to-Cradle PCF. This paper provides guidance and insights to help engineers develop effective strategies for material selection
Halsband, AdamLeinemann, TomkeBeer, MarkusHaiss, Eric
To provide an affordable and practical platform for evaluating driving safety, this project developed and assessed 2 enhancements to an Unreal-based driving simulator to improve realism. The current setup uses a 6x6 military truck from the Epic Games store, driving through a pre-designed virtual world. To improve auditory realism, sound cues such as engine RPM, braking, and collision sounds were implemented through Unreal Engine's Blueprint system. Engine sounds were dynamically created by blending 3 distinct RPM-based sound clips, which increased in volume and complexity as vehicle speed rose. For haptic feedback, the road surface beneath each tire was detected, and Unreal Engine Blueprints generated steering wheel feedback signals proportional to road roughness. These modifications were straightforward to implement. They are described in detail so that others can implement them readily. A pilot study was conducted with 3 subjects, each driving a specific route composed of a straight
Duan, LingboXu, BoyuGreen, Paul
Topology reasoning plays a crucial role in understanding complex driving scenarios and facilitating downstream planning, yet the process of perception is inevitably affected by weather, traffic obstacles and worn lane markings on road surface. Combine pre-produced High-definition maps (HDMaps), and other type of map information to the perception network can effectively enhance perception robustness, but this on-line fused information often requires a real-time connection to website servers. We are exploring the possibility to compress the information of offline maps into a network model and integrate it with the existing perception model. We designed a topology prediction module based on graph attention neural network and an information fusion module based on ensemble learning. The module, which was pre-trained on offline high-precision map data, when used online, inputs the structured road element information output by the existing perception module to output the road topology, and
Kuang, QuanyuRui, ZhangZhang, SongYixuan, Gao
In this paper, an incremental coordinated control method through anti-squat/lift/dive suspension is proposed based on and suited to a distributed drive electric vehicle with front and rear dual motors. The precise relationship between the suspension reaction force and the driving force of the wheel is derived as the control model through an in-depth analysis of the wheel motion and force. Through imposing the first-order dynamics, the proposed method not only provides the longitudinal speed control of the vehicle but also suppresses the longitudinal, vertical and pitch vibration of the vehicle. Simulation results show that the suspension reaction force formula derived in this paper is more suitable for dynamic conditions, and compared with the control method based on the simplified suspension anti-squat/lift/dive control model, the proposed method using the accurate control model has superior comprehensive control performance.
Feng, CongWu, GuangqiangYang, Yuchen
Automated driving is an important development direction of the current automotive industry. Level 3 automated driving allows the driver to perform non-driving related tasks (NDRTs) during automated driving, however, once the operating conditions exceed the designed operating domain, the driver is still required to take over. Therefore, it is important to rationally design takeover requests (TORs) in Level 3 conditional automated driving. This paper investigates the effect of directional tactile guidance on driver takeover performance in emergency obstacle avoidance scenarios during the transfer of control from automated driving mode to manual driving. 18 participants drove a Level 3 conditional automated driving vehicle in a driving simulator on a two-way four-lane urban road, performed a takeover, and avoided obstacles while performing non-driving related tasks. The driver's takeover performance during the takeover process was measured and subjective driver evaluation data was
Liang, XinyingLiang, YunhanMa, XiaoyuanWang, LuyaoChen, GuoyingHu, Hongyu
Drone show accidents highlight the challenges of maintaining safety in what engineers call “multiagent systems” — systems of multiple coordinated, collaborative, and computer-programmed agents, such as robots, drones, and self-driving cars.
This paper presents a new regression model-based method for accurate predictions of stiffness of different glass laminate constructions with a point-load bending test setup. Numerical FEA models have been developed and validated with experimental data, then used to provide training data required for the statistical model. The multi-variable regression method considered six input variables of total glass thickness, thickness ratio of glass plies as well as high-order terms. Highly asymmetrical, hybrid laminates combining a relatively thick soda-lime glass (SLG) ply joined with a relatively thin Corning® Gorilla® Glass (GG) ply were analyzed and compared to standard symmetrical SLG-SLG constructions or a monolithic SLG with the same total glass thickness. Both stiffness of the asymmetrical laminates and the improvement percentage over the standard symmetrical design can be predicted through the model with high precision.
Yu, ChaoCleary, ThomasJoubaud, Laurentkister, EvanFisher, W Keith
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
In the automotive industry, there have been many efforts of late in using Machine Learning tools to aid crash virtual simulations and further decrease product development time and cost. As the simulation world grapples with how best to incorporate ML techniques, two main challenges are evident. There is the risk of giving flawed recommendations to the design engineer if the training data has some suspect data. In addition, the complexity of porting simulation data back and forth to a Machine Learning software can make the process cumbersome for the average CAE engineer to set up and execute a ML project. We would like to put forth a ML workflow/platform that a typical CAE engineer can use to create training data, train a PINN (Physics Informed Neural Network) ML model and use it to predict, optimize and even synthesize for any given crash problem. The key enabler is the use of an industry first data structure named mwplot that can store diverse types of training data - scalars, vectors
Krishnan, Radha
Drivers sometimes operate the accelerator pedal instead of the brake pedal due to driver error, which can potentially result in serious accidents. To address this, the Acceleration Control for Pedal Error (ACPE) system has been developed. This system detects such errors and controls vehicle acceleration to prevent these incidents. The United Nations is already considering regulations for this technology. This ACPE system is designed to operate at low speeds, from vehicle standstill to creep driving. However, if the system can detect errors based on the driver's operation of the accelerator pedal at various driving speeds, the system will be even more effective in terms of safety. The activation threshold of ACPE is designed to detect operational errors, and it is necessary to prevent the system from being activated during operational operations other than operational errors, i.e., false activation. This study focuses on the pedal operation characteristics of pedal stroke speed and
Natsume, HayatoShen, ShuncongHirose, Toshiya
This literature review examines the concept of Fitness to Drive (FTD) and its impairment due to drug consumption. Using a Systematic Literature Review (SLR) methodology, the paper analyzes literature from mechanical engineering and related fields to develop a multidisciplinary understanding of FTD. Firstly, the literature is analysed to provide a definition of FTD and collect methods to assess it. Secondly, the impact of drug use on driving performance is emphasized. Finally, driving simulators are presented as a valid possibility for analysing such effects in a safe, controlled and replicable environment. Key findings reveal a lack of a comprehensive taxonomy for FTD, with various assessment protocols in use. Only static simulators are employed for drug evaluation, limiting realism and result reliability. Standard Deviation of Lane Position (SDLP) emerges as a gold-standard measure for assessing driver performance. Future research should focus on developing standard definitions for
Uccello, LorenzoNobili, AlessandroPasina, LucaNovella, AlessioElli, ChiaraMastinu, Gianpiero
The ISO TR 5469 Technical Report provides a framework to classify the AI/ML technology based on usage level and the properties and requirements to mitigate cyber and functional safety risks for the technology. This paper provides an overview of the approach used by ISO TR 5469 as well as an example of how one of the six ISO TR 5469 desirable properties (resilience to adversarial and intentional malicious input) can be analyzed for adversarial attacks. This paper will also show how a vehicle testbed can be used to provide a student with an AI model that can be used to simulate a non-targeted cyber security attack. The testbed can be used to simulate a poisoning attack where the student can manipulate a training data set to deceive the AI model during a simulated deployment.1 The University of Detroit Mercy (UDM) has developed Cyber-security Labs as a Service (CLaaS) to support teaching students how to understand and mitigate cyber security attacks. The UDM Vehicle Cyber Engineering (VCE
Zachos, MarkSeifert, Heinz
This paper describes a novel invention which is an Intrusion Detection System based on fingerprints of the CAN bus analogue features. Clusters of CAN message analogue signatures can be associated with each ECU on the network. During a learning mode of operation, fingerprints can be learnt with the prior knowledge of which CAN identifier should be transmitted by each ECU. During normal operation, if the fingerprint of analogue features of a particular CAN identifier does not match the one that was learnt then there is a strong possibility that this particular CAN identifier’s message is symptomatic of a problem. It could be that the message has been sent by either an intruder ECU or an existing ECU has been hacked to send the message. In this case an intruder can be defined as a device that has been added to the CAN bus OR a device that has been hacked/manipulated to send CAN messages that it was not designed to (i.e. could be originally transmitted by another device). It could also be
Quigley, ChristopherCharles, David
The research activity aims at defining specific Operational Design Domains (ODDs) representative of Italian traffic environments. The paper focuses on the human-machine interaction in Automated Driving (AD), with a focus on take-over scenarios. The study, part of the European/Italian project “Interaction of Humans with Level 4 AVs in an Italian Environment - HL4IT”, describes suitable methods to investigate the effect of the Take-Over Request (TOR) on the human driver’s psychophysiological response. The DriSMI dynamic driving simulator at Politecnico di Milano has been used to analyse three different take-over situations. Participants are required to regain control of the vehicle, after a take-over request, and to navigate through a urban, suburban and highway scenario. The psychophysiological characterization of the drivers, through psychological questionnaires and physiological measures, allows for analyzing human factors in automated vehicles interactions and for contributing to
Gobbi, MassimilianoBoscaro, LindaDe Guglielmo, VeronicaFossati, AndreaGalbiati, AndreaMastinu, LedaPonti, MarcoMastinu, GianpieroPreviati, GiorgioSabbioni, EdoardoSignorini, Maria GabriellaSomma, AntonellaSubitoni, LucaUccello, Lorenzo
Drivers present diverse landscapes with their distinct personalities, preferences, and driving habits influenced by many factors. Though drivers' behavior is highly variable, they can exhibit clear patterns that make sorting them into one category or another possible. Discrete segmentation provides an effective way to categorize and address the differences in driving style. The segmentation approach offers many benefits, including simplification, measurement, proven methodology, customization, and safety. Numerous studies have investigated driving style classification using real-world vehicle data. These studies employed various methods to identify and categorize distinct driving patterns, including naturalist differences in driving and field operational tests. This paper presents a novel hybrid approach for segmenting driver behavior based on their driving patterns. We leverage vehicle acceleration data to create granular driver segments by combining event and trip-based methodologies
Chavan, Shakti PradeepChinnam, Ratna Babu
To alleviate the problem of reduced traffic efficiency caused by the mixed flow of heterogeneous vehicles, including autonomous and human-driven vehicles, this article proposes a vehicle-to-vehicle collaborative control strategy for a dedicated lane in a connected and automated vehicle system. First, the dedicated lane’s operating efficiency and formation performance are described. Then, the characteristics of connected vehicle formations are determined, and a control strategy for heterogeneous vehicle formations was developed. Subsequently, an interactive strategy was established for queueing under the coordination of connected human-driven and autonomous vehicles, and the queue formation, merging, and splitting processes are divided according to the cooperative interaction strategy. Finally, the proposed lane management and formation strategies are verified using the SUMO+Veins simulation software. The simulation results show that the dedicated lane for connected vehicles can
Zhang, XiqiaoCui, LeqiYang, LonghaiWang, Gang
Scenario-based testing has become one of the important elements to evaluate the performance of automated vehicle systems before deploying on actual road. There are several approaches that can be used to conduct scenario-based testing via simulation approach. One of the important aspects in scenario-based safety testing is the driver-in-the-loop (DiL) simulation where it involves integration of hardware and human interaction. Therefore, motion platform-based vehicle driving simulators are commonly used for the DiL simulation for scenario-based testing. Generally, a high degree of freedom driving simulator is used for scenario-based testing such as 6 degrees of freedom (DoF) to achieve high accuracy to represent an actual vehicle response. Moreover, most of the motion platforms are designed using hexapod configuration, which also contributes to 6-DoF. However, this type of design requires large space to conduct the testing because the field of motion (FoM) is high in three axes and high
Kleolee, KahOnnAparow, Vimal RauCheok, Jun Hongde Boer, NielsJamaluddin, Hishamuddin
Having an in-depth comprehension of the variables that impact traffic is essential for guaranteeing the safety of all drivers and their automobiles. This means avoiding multiple types of accidents, particularly rollover accidents, that may have the capacity of causing terrible repercussions. The non-measured factors in the system state can be estimated employing a vehicle model incorporating an unknown input functional observer, this gives an accurate estimation of the unknown inputs such as the road profile. The goal of the proposed functional observer design constraints is to reduce the error of estimation converging to a value of zero, which results in an improved calculation of the observer parameters. This is accomplished by resolving linear matrix inequalities (LMIs) and employing Lyapunov–Krasovskii stability theory with convergence conditions. A simulator that enables a precise evaluation of environmental factors and fluctuating road conditions was additionally utilized. This
Saber, MohamedOuahi, MohamedNaami, GhaliEl Akchioui, Nabil
Handling and ride comfort optimization are key vehicle design challenges. To analyze vehicle performance and investigate the dynamics of the vehicle and its subcomponents, we rely heavily on robust experimental data. The current article proposes an outdoor cleat test methodology to characterize tire dynamics. Compared to indoor procedures, it provides an effective tire operating environment, including the suspensions and the vehicle chassis motion influence. In addition, it overcomes the main limitation of existing outdoor procedures, the need for dedicated cleat test tracks, by using a set of removable cleats of different sizes. A passenger vehicle was equipped with sensors including an inertial measurement unit, a noncontact vehicle speed sensor, and a wheel force transducer, providing a setup suitable to perform both a handling test routine and the designed cleat procedure, aimed at ride testing and analysis. Thus, the outdoor cleat test data were compared with indoor test
Gravante, GerardoNapolitano Dell’Annunziata, GuidoBarbaro, MarioFarroni, Flavio
For years, Proffesor Bozhi Tian’s lab has been learning how to integrate the world of electronics — rigid, metallic, bulky — with the world of the body — soft, flexible, delicate.
Technology for lane line semantic segmentation is crucial for ensuring the safe operation of intelligent cars. Intelligent cars can now comprehend the distribution and meaning of scenes in an image more precisely thanks to semantic segmentation, which calls for a certain degree of accuracy and real-time network performance. A lightweight module is selected, and two previous models are improved and fused to create the lane line detection model. Finally, experiments are conducted to confirm the model's efficacy. This paper proposes a lightweight replacement program with the aim of addressing the issue of large parameterization in the generative adversarial network (GAN) model and difficult training convergence. The overall network structure is selected from the Pix2Pix network in the conditional generative adversarial network, and the U-net network of the generator is cut and replaced by the Ghost Module, which consists of a modified downsampling module that enhances the global fusion
Yang, KunWang, Jian
Nowadays, cognitive distraction in the process of driving has become a frequent phenomenon, which has led to a certain proportion of traffic accidents, causing a lot of property losses and casualties. Since the fact that cognitive distraction is mostly reflected in the driver's reception and thinking of information unrelated to driving, it is difficult to recognize it from the driver's facial features. As a result, the accuracy of prediction is usually lower relying solely on facial performance to detect cognitive distraction. In this research, fifty participants took part in our simulated driving experiment. And each participant conducted the experiment in four different traffic scenarios using a high-fidelity driving simulator, including three cognitive distraction scenarios and one normal driving scenarios. Firstly, we identified the facial performance indicators and vehicle performance indicators that had a significant effect on cognitive distraction through one-way ANOVA. Then we
Qu, ChixiongBao, QiongQu, QikaiShen, Yongjun
This study tackles the issue of order delays in logistics using XGBoost for feature analysis and reinforcement learning for intelligent courier scheduling. Pickup order data from May 1 to October 31, 2023, in Chongqing is analyzed using spatio-temporal statistical methods. Key findings include that order placement peaks at 9:00 a.m., delays peak at 10:00 a.m., and the delay rate is 8.6%. A significant imbalance exists between the regional daily average of dispatchable couriers and order volumes.XGBoost is employed to predict order delays, revealing that pickup location is the most influential factor (27%), followed by courier pickup location (22%). These factors and their relationships are identified as key drivers of delays.To address these issues, a reinforcement learning-based courier scheduling optimization model is developed. The model defines courier location, current time, and pending orders as state variables and adopts an epsilon-greedy strategy for action selection
Wang, ManjunYu, Xinlian
Developing models for predicting the low-temperature cracking resistance of asphalt mixtures is a complex process with a wide variety and complex influence mechanisms of variables, leading to higher uncertainty in the prediction results. Several models have been developed in this regard. This study developed a Bayesian neural network (BNN) model for predicting the fracture energy of low-temperature semi-circular bending (SCB) tests based on pavement condition measurements, traffic, climate, and basic parameters of the material. The model was trained and evaluated using low-temperature SCB test data from in-situ pavement core samples, and the results showed that the coefficient of determination (R2) of the BNN model was greater than 0.8 for both the training and testing sets. The variable importance scores showed that the decrease of transverse crack rating index (TCEI) and gradation were the most important factor affecting low-temperature fracture energy and that the ambient
Song, ZiyuNi, FujianHuang, JiaqiJiang, Jiwang
Tunnel linings are an important safeguard for the integrity and stability of tunnels. However, cracks in the tunnel lining may have extremely unfavourable consequences. With the acceleration of urbanisation and the increasing construction of tunnels, the problem of cracks in the concrete lining is becoming more and more prominent. These cracks not only seriously affect the stability of the structure, but also pose a serious threat to the safety of tunnel operation. If left unchecked, the cracks may expand further and cause various safety hazards, such as water leakage and falling blocks. This in turn will undermine the normal function of the tunnel and endanger the lives of tunnel users. It has been proved that the traditional manual method of detecting cracks in tunnels has problems such as low accuracy and low efficiency. In order to solve this problem, it is very necessary for this study to pioneer an intelligent method for identifying tunnel lining cracks using the YOLOv11
Zhang, YalinNiu, PeiGuo, FengYan, WeiLiu, JianKou, Lei
The performance differences of multiple sensors lead to inconsistencies, incompleteness, and distortion in the perception data of multi-source vehicle information in highway scenarios. Optimizing data fusion methods is important for intelligent toll collection systems on highways. First, this paper constructs a dataset for matching and fusing multi-source vehicle information in highway gantry scenarios. Second, it develops convolutional neural network models, Match-Pyramid-MVIMF-EGS and CDSSM-MVIMF-EGS, for this purpose. Finally, comparative experiments are conducted based on the constructed dataset to assess the performance of the Match-Pyramid-MVIMF-EGS and CDSSM-MVIMF-EGS models. The experimental results indicate that the Match-Pyramid-MVIMF-EGS model performs better than the CDSSM-MVIMF-EGS model, achieving matching and fusion accuracy of 93.07%, precision of 95.71%, recall of 89.17%, F1 scores of 92.32%, and 186 of training throughput respectively.
Wang, JunjunZhao, Chihang
The introduction of autonomous vehicles (AVs) promises significant improvements to road safety and traffic congestion. However, mixed-autonomy traffic remains a major challenge as AVs are ill-suited to cooperate with human drivers in complex scenarios like intersection navigation. Specifically, human drivers use social cooperation and cues to navigate intersections while AVs rely on conservative driving behaviors that can lead to rear-end collisions, frustration from other road users, and inefficient travel. Using a virtual driving simulator, this study investigates the use of a human factors-informed cooperation model to reduce AV reliance on conservative driving behaviors. Four intersection scenarios, each involving a left-turning AV and a human driver proceeding straight, were designed to obfuscate the right-of-way. The classification models were trained to predict the future priority-taking behavior of the human driver. Results indicate that AVs employing the human factors-informed
Ziraldo, ErikaOliver, Michele
Noise, Vibration, and Harshness (NVH) simulations of vehicle bodies are crucial for assessing performance during the design phase. However, these simulations typically require detailed computer-aided design (CAD) models and are time-consuming. In the early stages of vehicle development, when only high-level vehicle sections are available, designing the body-in-white (BIW) structure to meet target values for bending and torsional stiffness is challenging and often requires multiple iterations. To address these challenges, this study deploys a reduced-order beam modelling approach. This method involves identifying the beam-like sections and major joints within the BIW and calculating their sectional properties (area, area moments of inertia along the plane’s independent axes, and torsion constant). These components form a simplified skeleton model of the BIW. Load and boundary conditions are applied to the suspension mount locations at the front and rear of the vehicle, and torsional and
Khan, Mohd Zishan AliThanapati, AlokDeshmukh, Chandrakant
At present, due to the complexity and nonlinearity, the thermal safety and economic feasibility assessment and optimization of the Solid Oxide Fuel Cell-Gas Turbine (SOFC-GT) system under variable loads is important to extend the service life and reduce the cost. To solve these problems, this paper proposes a top-level cyclic SOFC-GT system, which considers the design of two-stage preheaters, as well as the impact of material reaction kinetics and thermoelectric coupling characteristics on system performance. Furthermore, the multi-criteria evaluation of the SOFC-GT system under variable loads has been studied, with evaluation indicators primarily including thermodynamic and economic indicators. Afterwards, a Spearman-based parametric sensitivity analysis is used to explore the response trends of performance indicators within the SOFC-GT system. Additionally, an intelligent learning method based on convolutional neural network is designed to determine the dynamic behavior between
Fan, LiyunKui, XuChen, ChenShen, ChongchongLi, BoWei, Yunpeng
The degradation of vehicle performance resulting from powertrain degradation throughout the lifecycle of alternative energy vehicles (AEVs) has consistently been a focal issue among scholars and consumers. The purpose of this paper is to utilize a one-dimensional vehicle simulation model to analyze the changes in power performance and economy of fuel cell vehicles as the Proton Exchange Membrane Fuel Cell (PEMFC) stack degrades. In this study, a simulation model was developed based on the design parameters and vehicle architecture of a 45kW fuel cell vehicle. The 1D model was validated for accuracy using experimental data. The results indicate that as the stack performance degrades, the attenuation rate of the fuel cell engine is further amplified, with a degradation of up to 13.6% in the system's peak output power at the End of Life (EOL) state after 5000 hours. Furthermore, the level of economic performance degradation of the complete vehicle in the EOL state is dependent on the
Li, YouDu, JingGuo, DonglaiWang, KaiWang, Yupeng
As a clean energy, low carbon and pollution-free, hydrogen is the preferred alternative fuel for traditional internal combustion engines. However, how to use hydrogen internal combustion engine to achieve satisfactory performance under vehicle conditions is still a challenge.In this paper, a vehicle simulation model is established based on a modified 25-ton hydrogen internal combustion engine truck, and the model is designed as a hybrid model by selecting a suitable motor. The two models are used to simulate the CHTC (China Heavy-duty Commercial Vehicle Test Cycle) cycle conditions. According to the simulation results, compared with the original vehicle's power performance and economy, the results show that the power performance is increased by 100%, and the economy is increased by 20%. Hybrid technology can effectively improve the performance of the vehicle.
Bai, Xueyan
Monitoring the rotor temperature of drive machines is crucial for the safety and performance of electric vehicles. However, due to the complex operating conditions of electric vehicles, the thermal parameters of vehicular induction machines (IMs) vary significantly and are difficult to identify accurately. This article first establishes a concise but effective thermal network for IMs and analyzes the influencing factors of thermal parameters. Then, a parameter identification network (PIN) with multiple parallel branches is constructed to learn the mapping relationship between electromechanical variables and thermal parameters. Afterward, temperature datasets for network training are built through bench testing. Finally, the effectiveness of identified parameters for rotor temperature estimation application is verified, demonstrating improved interpretability, generalization ability, and accuracy compared to an end-to-end neural network.
Jiang, ShangHu, Zhishuo
The increased popularity of electric vehicles featuring distributed powertrains is enabling an easy and cost-effective implementation of torque vectoring. This is a renowned technique for controlling vehicle lateral dynamics having the objective of improving both vehicle handling and stability. Nevertheless, the application of torque vectoring at the front axle can increase the difficulty of usual driving tasks. This is because differential longitudinal forces at front tires generate a steering wheel torque, which can be badly perceived by the driver, up to the point of jeopardizing the benefits of having a torque vectoring control. The aim of this article is thus to study in detail the steering torque corruption caused by front axle torque vectoring for proposing some electric power steering control strategies compensating for this effect. Indeed, the electric power steering controllers developed in this study are designed based on the analytical derivation of the torque steer theory
Asperti, MicheleVignati, MicheleSabbioni, Edoardo
Items per page:
1 – 50 of 6264