Browse Topic: Data acquisition and handling

Items (6,184)
Find
An Efficient Approach to Key Negotiation in an Automotive Environment2026-01-0087To be published on 04/07/2026
Negotiating Keys for applications such as message authentication within a vehicle presents many problems as, in designing the algorithm; the algorithm must be able to be utilized by small, fixed-point processors. In addition, if there is a desire to do this algorithm in the manufacturing environment, there are severe time constraints placed on how long this algorithm can take, as there are strict station time requirements, which are expensive to change, and any time utilized in the plant can negatively affect vehicle throughput. Additionally, negotiating these keys between many ECUs can greatly increase the time required to negotiate a common key using standard multi-party Diffie-Hellman. Timing would also be an issue in the case of using pair-wise Diffie-Hellman for encryption and distribution of keys utilizing a key master. To solve these problems in multi-party key negotiation, we have utilized the Elliptic Curve variation of the Burmester-Desmedt (ECBD) algorithm. ECBD is
Van Dam, Theo, Mazzara, Bill
Optimal Vehicle Sampling for Fleet-Wide Emissions Monitoring using Submodular Maximization2026-01-0283To be published on 04/07/2026
This article addresses the problem of optimal vehicle sampling for fleet-wide in-use emissions monitoring, a necessity driven by the absence of direct emissions sensors in modern production vehicles and the variable impact of in-use changes and operational factors (mileage, time-in-service, workload) on emissions performance across a fleet. Recognizing that comprehensive fleet testing is impractical due to significant downtime and cost, we propose a novel approach to identify a small, yet optimally informative subset of vehicles for sampling. The proposed approach leverages submodular function maximization, a technique rooted in optimal experimental design, specifically D-optimal design, to maximize the determinant of the information matrix (e.g., of $X^TX$, where $X$ is the regressor/design matrix in the case of a linear in parameters model). This approach ensures that the collected data yields maximum information for refining and building accurate models for emissions characteristic
ZHANG, JIADI, Li, Xiao, Kolmanovsky, Ilya, Tsutsumi, Munecika, Nakada, Hayato
Understanding Adversarial Transferability in Vision-Language Models for Autonomous Driving: A Cross-Architecture Analysis2026-01-0170To be published on 04/07/2026
Vision-language models (VLMs) offer a promising end-to-end approach for autonomous driving, but their robustness to physical adversarial attacks remains an unaddressed gap for safety-critical deployment. This paper introduces a systematic framework for evaluating these attacks using the Dolphins VLM and CARLA simulator. We employ Natural Evolution Strategies (NES), a black-box optimization method, to generate adversarial patches without internal model access. Our approach uses semantic similarity loss and Expectation over Transformation (EoT) to ensure physical realizability under varying viewing conditions. Patches are strategically placed on realistic advertising infrastructure (bus shelter panels and highway billboards), reflecting a plausible threat model. We establish two complementary scenarios: crosswalk pedestrian detection suppression and highway steering manipulation. Our evaluation reveals critical vulnerabilities. Adversarial patches achieved a 76.0% overall attack success
Fernandez, David, MohajerAnsari, Pedram, Salarpour, Amir, Pese, Mert D.
LiDAR-Driven A* Path Planning with Dynamic Obstacle Avoidance for Autonomous Navigation in ROS2026-01-0030To be published on 04/07/2026
Autonomous mobile robots are becoming a key part of everyday operations in industries like manufacturing, logistics, healthcare, and even home assistance. A core requirement for these robots is the ability to navigate efficiently and reliably within their operating environments. To do this automation, the robot needs to understand its surroundings, figure out where it is on a map, and find a safe path from where it is to where it needs to go without bumping into anything. This paper presents an effective grid-based path planning solution for autonomous indoor navigation with a mobile robot. Achieving reliable and collision-free navigation in changing environments is a major challenge for mobile robotics. This is especially true when obstacles can appear unexpectedly, requiring quick re-planning. To tackle this issue, an improved A* algorithm was implemented to work closely with LiDAR for environmental awareness. The improved algorithm was added to the robot’s navigation system, and
Devaraj, Sriram Sanjeev, Park, Jungme
Numerical Analysis of Operating Parameters Coupled Effects on Flow Instability Mechanisms at Supercritical Pressure2026-01-0134To be published on 04/07/2026
Supercritical fluids (SCFs) are widely utilized in advanced energy systems due to their exceptional heat transfer characteristics. However, flow instability, particularly near pseudo-critical conditions, poses a significant challenge to their effective applications. Despite extensive research, the coupled effects of thermal and fluid dynamic mechanisms on flow instability remain insufficiently understood, particularly in supercritical water (SCW) systems. This study identifies the onset of flow instability (OFI) in the negative slope region of pressure drop curve during steady-state conditions, analyzing the relationship between pressure drop and mass flow rate. Transient analyses were then conducted to investigate the influence of operating parameters on oscillatory phenomena such as pressure drop oscillations (PDO), density wave oscillations (DWO), and thermal oscillations (ThO). The results show that higher mass flow rates and heat fluxes, reduce pressure oscillations by up to 20
Rehman, Attiq Ur, Ahmad, Hassaan, Du, Xiaocheng, Kanwal, Lubna
Ensemble of Experts for Generating Digital Parking Maps from Remote Sensing Images2026-01-0107To be published on 04/07/2026
Being a vital component of transportation, parking and parking-related research are gaining increasing attention from the intelligent transportation and automotive industry research communities. A digital parking map with precise parking spot geospatial information is crucial for research such as automatic valet parking, parking spot recommendations, and parking route optimization. Currently, these maps are generated through costly and resource-intensive human annotation. To address the challenges of insufficient parking maps and the high costs of manual map creation, researchers have proposed the following two strategies for generating parking maps. The SLAM-based map generation for indoor parking lots and the remote sensing image-based map generation for outdoor parking lot. Our work mainly focuses on the outdoor high-definition parking map generation from remote sensing images. These images often suffer from occlusion, inconsistent resolution, and varying luminosity conditions. To
Shukla, Ajitesh, Cao, Xiaofei, Liu, Yongkang, Takeuchi, Yusuke, Sisbot, Akin
The New Era for Automotive Cybersecurity An Opportunity for Standardization of Automotive Cybersecurity2026-01-0086To be published on 04/07/2026
The automotive industry is moving from a reactive, self-determined approach to cybersecurity to a standardized, regulated one. This document, an update on collaborative work from the SAE TEVEES18B Committee and GlobalPlatform Automotive Task Force, outlines this transition. In the past, cybersecurity challenges in the automotive industry were seen as novel, and each company handled them individually. This era was characterized by self-determination of security mitigations and a unique justification for solutions. Companies often developed cybersecurity concepts based on past incidents and their own risk tolerance. A key tool during this time was the TARA (Threat Analysis and Risk Assessment), which organizations used to justify their security concepts or lack thereof. This led to company-specific policies and a lack of a global standard. The result was a fragmented market where every new project required a novel cybersecurity concept, hindering the development of a robust market for
Mazzara, Bill, Rawlings, Craig
Implementation and evaluation of a method to defend in-vehicle networks using SAE J1939-91C to authenticate network messages between multiple ECUs.2026-01-0092To be published on 04/07/2026
The objective of this paper is to understand the effort required to integrate the hardware and software of in-vehicle cybersecurity systems. The in-vehicle cybersecurity method discussed is the SAE J1939-91C, which involves Network formation, Rekeying, and secure Message Exchange between Electronic Control Units (ECUs). The SAE J1939-91C network security protocol operates over a CAN-FD network to perform necessary cryptographic operations and key generation. To evaluate the method, test vectors were created to validate SAE J1939-91C key generations and cryptographic operations on the simulated ECU in-vehicle network system hardware. The evaluation results provide a benchmark for network system performance that can be used in a system test bench to assess compliance with the SAE J1939-91C standard and to test the standard's effectiveness in ambiguous edge cases. In the future, this benchmark could be utilized in "Plugtests" to assign test scores to non-uniform SAE J1939-91C
Zachos, Mark, Medam, Krishna Teja
Optimization Strategies of Gliding Arc Plasma-Based Ammonia Cracker towards On-Board Applications2026-01-0292To be published on 04/07/2026
Ammonia has emerged as a viable hydrogen energy carrier owing to its superior hydrogen density, mature industrial utilization. However, ammonia faces critical challenges including inadequate ignition characteristics and sluggish combustion kinetics, necessitating supplementary high-reactivity fuels for optimizing combustion. Onboard ammonia cracking technology resolves this problem through on-demand hydrogen in-time production. Among existing decomposition methods, gliding arc plasma (GAP) demonstrates exceptional promise for on-board crackers given its decent hydrogen conversion rate and transient response capability. Prevailing research predominantly employs experimental methodologies wherein macroscopic parameters inadequately proxy microscopic electric field characteristics, thereby impeding mechanistic insight. This study establishes a quasi-1D numerical model to interrogate how electric field parameters modulate ammonia-to-hydrogen conversion yield and system energy efficiency in
Dong, Guangyu, LI, Xian, ZHOU, Yanxiong, XU, Jie, Li, Liguang
A Cooperative AFS/DYC Control for FWDIEV Based on Dissipative Energy Method2026-01-0619To be published on 04/07/2026
To enhance the lateral stability of four-wheel-drive intelligent electric vehicles (FWDIEV) under extreme operating conditions, this paper proposes a cooperative control strategy integrating active front steering (AFS) and direct yaw moment control (DYC) based on dissipative energy method. A nonlinear three-degree-of-freedom vehicle model is established to analyze the evolution of the vehicle state phase trajectory. A quantitative lateral stability index is constructed using dissipative energy to accurately evaluate the vehicle's lateral dynamics. Utilizing dissipative energy and its gradient information, a time-varying stability boundary is defined under dynamic constraints, and adaptive weighting coordination between the AFS and DYC systems is designed to achieve coordinated control of front steering angle and additional yaw moment. A feedforward–model predictive control (FF-MPC) framework is developed, in which a feedforward module generates compensation based on driver intent to
Zhao, Kun, Zhao, Zhiguo, Wang, Yutao, Xia, Xue, Chen, Xi, Hu, Yingjia
Collaborative Cybersecurity in a Shifting Automotive Supply Chain2026-01-0084To be published on 04/07/2026
The evolution toward software-defined vehicles (SDVs) is causing disruption to the traditional automotive supply chain and breaking down the common hierarchical OEM, tier 1 supplier, and tier 2 supplier relationships. With demands for faster software release cycles, more advanced software projects involving multi-party development, and considerations for end-to-end embedded and cloud integrations, new cybersecurity challenges are introduced that no single organization can address alone. Thus, this disruption creates new trust dependencies and requires new models for collaboration, transparency, and joint responsibility in cybersecurity. This paper presents a collaborative cybersecurity model, emphasizing shared responsibility during multi-party development between OEMs, tier 1 and 2 suppliers, engineering services organizations, and technology and services providers. As such, we explore collaborative approaches for each stage in the development lifecycle including design, development
Oka, Dennis Kengo, Vinzenz, Nico
RRT*_Heuristic_Adaptiv_maxR: A practical path planning method based on novel adaptive framework for autonomous mobile agent2026-01-0038To be published on 04/07/2026
Aim at addressing the problem of random sampling and low efficiency of rapidly-exploring random tree (RRT) path planning algorithm, a practical path planning method is proposed for autonomous mobile agents, named RRT*_Heuristic_Adaptiv_maxR. This method is based on a new adaptive framework, balancing the search efficiency and convergence accuracy of path planning by dynamically adjusting the maximum radius parameters in the search process. Combining the heuristic search strategy, guiding the sampling direction, accelerating convergence to the target area, the obstacle detection mechanism is used to optimize the path in real time to avoid collisions. Experimental results show that the proposed method, RRT*_Heuristic_Adaptiv_maxR, has higher planning success rate and path quality in complex environments compared with various RRT algorithms, and it is verified the effectiveness of the adaptive framework and the maximum radius dynamic adjustment strategy in mobile agent path planning.
Dong, Xiaomei, Zeng, Dequan, Petitti, Antonio, You, Jingyi, xiong, min, GAO, Letian, Jiang, Xiaomeng, Yu, Yinquan, CARBONE, GIUSEPPE, Hu, Yiming, Nawaz, Aamir, Linh, Nguyen Vu
Oscillatory Behavior of Understeering Vehicles2026-01-0623To be published on 04/07/2026
Oscillations in understeering vehicles are occasionally described in the literature, primarily in terms of the poles of the yaw rate response, but perhaps not completely appreciated in their complexity. This work shows that as speeds of the understeering vehicle increase, the increasingly underdamped poles of the yaw rate transfer function combine with the effects of a low frequency zero and a reduced steady-state response to result in oscillations greater than would be expected from eigenvalues alone. A speed range for acceptable yaw rate response is suggested, and it is shown that a typical understeering passenger car operates within this range. As the understeering vehicle’s speed increases beyond this range, the high speed limit of the oscillation frequency is found.
Williams, Daniel
A Control Method for Dual Motor Redundant Steer System based on Zeroing Neural Networks2026-01-0581To be published on 04/07/2026
In the process of driving a vehicle, the steering system directly determines the vehicle's travel path and direction, serving as the physical basis for precise control in autonomous driving. Whether it is lane keeping, turning, or obstacle avoidance, high responsiveness of the steering control is required to execute the commands of the decision-making system. This study focuses on the dual-motor steer-by-wire (SBW) system, addressing the tracking issues arising from factors such as dual-motor coupling. A control strategy for coordinated dual motor drive in an autonomous driving steering system based on dual-motor redundancy is proposed. A dynamic model of the steer-by-wire system is constructed, and the software algorithm employing a zeroing neural network for the dual-motor redundant steering is used to solve the synchronization issues of the dual-motor system, based on the concept of cross-coupled control. Finally, simulation analysis was performed on the MATLAB simulation platform
Yang, Lingang, xiong, min, Zeng, Dequan, Petitti, Antonio, Nawaz, Aamir, Linh, Nguyen Vu, Hu, Yiming, Yu, Yinquan, CARBONE, GIUSEPPE, GAO, LETIAN, Jiang, Xiaomeng
Mitigation of Sensor-Level Cyber-Physical Attacks on Electric Vehicle Powertrains: A Simulation-Based Study2026-01-0091To be published on 04/07/2026
Electric vehicles (EVs) rely extensively on sensor feedback for safe and efficient powertrain operation. However, this dependency introduces cyber-physical vulnerabilities, especially when sensor signals are maliciously manipulated. This paper presents a simulation-based investigation into sensor-level cyberattacks on a mid-sized EV powertrain model developed in MATLAB/Simulink. The study quantifies mechanical consequences and evaluates mitigation strategies to enhance system resilience. Four representative attack scenarios were simulated. Speed sensor spoofing led the controller to misinterpret vehicle velocity, causing a 41% overshoot beyond the 50 km/h setpoint. False data injection into torque/current sensors triggered an unintended torque surge of approximately 20%, resulting in inverter current saturation within 2 seconds. Battery temperature spoofing delayed thermal protection, allowing a deviation of 1.5 °C/min beyond safe operating limits. A hybrid attack combining frozen
Tariq, Usama, Sahandabadi, Sahereh, Dianat, Ali
Development of Tilt Test Facility2026-01-0265To be published on 04/07/2026
Mobility, fire power and protection are the three most important characteristics of an Armoured Fighting Vehicle (AFV). Transmission system facilitates mobility to the fighting vehicle. Combat Vehicles Research and Development Establishment (CVRDE) is developing Automatic Transmission for Armoured Fighting vehicles which includes gearbox, steering system, braking system and control system. In the battle field, AFVs have to operate at different gradients & side slopes. The Automatic Transmission system being developed by CVRDE is a hydro-mechanical automatic transmission in which hydraulic oil is used for lubrication & control of various systems of it. A Tilt Test Facility (TTF) has been designed, developed, and installed at CVRDE to simulate different angles of gradients from +35° to - 35° & side slopes from +20° to -20°. The TTF enables to evaluate the functional aspects of the transmission system at different angles. The TTF consists of two tiltable platforms, Rotary actuators
V, Uma, Naik, Sheshu
Impact of C-rates on Electrical Behavior of Li-ion Batteries under Coupled Temperature-load Conditions2026-01-0383To be published on 04/07/2026
The utilization of lithium-ion batteries in electric vehicles becomes increasingly common. Battery performance degradation is influenced by both temperature and boundary constraints. Monitoring the usable cell capacity (UCC) under real-world operating conditions is essential for ensuring accurate state-of-health estimation. Capacity tests in laboratory settings are typically conducted at low C-rates to approximate equilibrium conditions, whereas in real vehicle applications, charging currents are often much higher. This discrepancy in charge/discharge rates frequently results in deviations between laboratory characterization and on-board BMS capacity estimation. To investigate the impact of C-rates under coupled temperature-load conditions, we conducted long-term cycling tests on LFP/graphite pouch cells at 25 °C and 45 °C under different constrained conditions. The UCC was evaluated at different test rates (1/20C, 1/3C, and 1C) as an indicator of aging. The results show that
ZHANG, Shan, Niu, Zhice, Xia, Yong
Development and Trial of a Non-Dominated Sorting Genetic Algorithm for Rapid Engine Calibration2026-01-0285To be published on 04/07/2026
In the quickly changing automotive powertrain landscape, calibration processes are a major bottleneck to production and new technology development. The implementation of rapid powertrain calibration processes can reduce development resource requirements, improve performance, and reduce risk. In previous works the author successfully developed an internal combustion engine calibration strategy using a Lagrange-Multiplier genetic algorithm for multi-constraint optimization. In the present work, this previous algorithm is expanded to create and demonstrate a new non-dominated sorting genetic algorithm which is aimed at identifying the entire Pareto front of the relationship between two engine performance parameters. Accurate and rapid identification of Pareto fronts enables designers to effectively and efficiently balance performance tradeoffs and identify optimal calibration settings. The algorithm developed here optimized the calibration of three combustion control parameters
Mansfield, Andrew
Energy Management Strategy for Hybrid Mining Trucks Based on Global SOC Planning and Neural Network Optimal Control2026-01-0446To be published on 04/07/2026
Hybrid mining trucks, as core equipment for mine transportation, face high energy consumption and significant fluctuations in power demand during cyclic operations due to prolonged exposure to demanding operating conditions characterized by heavy loads and variable working conditions. To address the issues of high energy consumption and significant fluctuations in power demand during the cyclic operation of mining trucks, this paper proposes a hybrid mining truck energy management strategy based on global SOC (State of Charge) planning and neural network optimization control. First, a powertrain model was developed for a typical operating cycle of a hybrid mining truck, and its accuracy was validated by comparing it with experimental data. Using dynamic programming algorithms to plan the SOC for single-cycle operations provides a rational reference for energy allocation across different operational phases of mining trucks during a single cycle. Next, using the powerful nonlinear
Yang, Jianyu, Zhao, Zhiguo, Chen, Huiyong, Li, Tao, ZHUANG, Wenyu, Shen, Peihong, Tang, Peng
An Image-Derived Multimodal Framework for Automotive Tire Exterior Noise Prediction2026-01-0220To be published on 04/07/2026
This study presents an image-derived multimodal AI framework for early-stage tire noise evaluation. The proposed model requires only multi-angle photographs captured by a standard smartphone and basic tire specifications. From these images, scaled three-dimensional (3D) meshes and fixed-view depth maps are reconstructed and combined with numerical parameters within a neural network architecture. Three input branches—a point-cloud–gradient branch, a depth-map convolutional neural network (CNN) branch, and a specification multi-layer perceptron (MLP) branch—are jointly trained using a composite loss that integrates frequency-weighted mean squared error (MSE), spectral cosine similarity, FFT-domain consistency, and A-weighted sound-level terms. A dataset of 28 tires, spanning passenger, SUV, and pickup applications for both battery electric vehicles (BEVs) and internal combustion engine (ICE) vehicles, was evaluated using leave-one-out (LOO) cross-validation. The model achieved a mean
Shao, Guangxin, Shopoff, Scott, Franklin, Nicholas
High Strain-Rate Behaviors of Fiber Reinforced Composites in Principal and Off-Axis Directions2026-01-0175To be published on 04/07/2026
Materials can exhibit significantly different mechanical behaviors compared to quasi-static conditions at high strain rates (> 100 s-1). High strain rate tests using setups such as SHPB (Split-Hopkinson Pressure Bar) can provide, in a practicable manner, the stress-strain relations for a material at high strain rates. Such properties are vitally needed for activities such as simulation-driven impact safety design of composite structures deployed in the form of automotive body parts and assembly, and other sub-systems. Although the behaviors of isotropic and ductile materials such as various metallic alloys appear to have been extensively studied and reported in literature, dependence of mechanical properties of fiber-reinforced composites especially in different off-axis directions are extremely difficult to come across. To fill up this void, a detailed experimental study has been carried out on high strain rate mechanical characterization of orthotropic glass, carbon and a natural
Bawa, Prashant, Deb, Anindya, Zhu, Feng
Neural Surrogates and Reinforcement Learning for Heavy-Duty Diesel Torque Control: A Data-Driven Alternative to PID2026-01-0199To be published on 04/07/2026
Accurately reproducing the torque trace of regulatory drive cycles is central to heavy-duty diesel emissions certification and repeatable development testing. Conventional controllers—PI/PID with gain scheduling and feedforward mapping—can meet requirements but demand extensive hand-tuning and struggle with strong nonlinearities and time-varying delay. Building on prior work that quantified the limits and benefits of gain-scheduled and feedforward strategies under the FTP transient cycle, this work investigates a fully data-driven control framework. First, a recurrent neural network (LSTM) torque surrogate is trained on dynamometer data to capture engine dynamics from readily available signals (e.g., engine speed, commanded pedal/torque, recent torque history) and actuator constraints. Second, this learned “world model” is used to train a reinforcement learning (RL) policy that outputs pedal commands to minimize torque-tracking error while regularizing control effort and respecting
Cook, James, Puzinauskas, Paulius, Bittle, Joshua, Hall, Spencer
Research on the control of magnetic weakening scheme for the voltage utilization rate of hybrid transmission2026-01-0447To be published on 04/07/2026
This paper proposes a weak magnetic control method for PMSM current regulators based on the ratio of pressure to frequency and voltage utilization rate when the motor speed exceeds the weak magnetic speed. That is, the direct-axis offset current id is generated through the action of weak magnetic feedforward and weak magnetic I regulator, and the control of weak magnetic speed can directly adjust the demagnetization effect of id to achieve weak magnetic control of PMSM. The weak magnetic control method of PMSM current regulators with the outer loop of voltage utilization rate is adopted. It is proposed to use the phase-locked loop control method to decode the motor position signal in a closed loop and calculate the motor speed. The fluctuation of the position signal is reduced and the signal is more stable. This paper realizes constant power speed regulation based on high voltage utilization rate and weak magnetic control method at high speed operation above the base speed, which
Jing, Junchao, Yan, Xiaoqian, Liu, Yiqiang, Sun, Fengjiao, Dai, Zhengxing
Adaptive Economic Model Predictive Control for Engine Emissions Management with Compensation for Performance Drift2026-01-0287To be published on 04/07/2026
This paper addresses the changes in engine emissions due to in-use component changes through the synergistic application of predictive control, machine learning, and onboard adaptation. In particular, we consider an adaptive economic Model Predictive Control (eMPC) strategy to mitigate the effects of performance drift on Nitrogen Oxides (NOx) and Soot emissions from compression ignition (diesel) engines. A performance drift block, which applies a multiplier and offset to nominal emissions, is integrated with a high-fidelity Neural Network (NN) plant model to simulate these characteristic changes. To counteract variability, two online adaptation methods are integrated within the eMPC framework: One is based on Recursive Least Squares (RLS) and another on a continuously updated online NN. The proposed control architecture is validated through simulations over standard transient cycles. Results demonstrate that while the rate-based eMPC possesses inherent robustness to performance drift
ZHANG, JIADI, Li, Xiao, Kolmanovsky, Ilya, Tsutsumi, Munecika, Nakada, Hayato
A Framework for Predicting 3D Printed Part Defects Using MultiModal Deep Learning on Infrared Images2026-01-0140To be published on 04/07/2026
Fused filament fabrication (FFF) has gained popularity in recent years because it can produce prototypes and functional components with complex geometry. Because of inherent process variability, the components often exhibit defects such as warping, layer delamination, voids, and poor surface finish, as well as issues related to variable material strength and anisotropy. In-situ monitoring (ISM) of the FFF process is a promising technique to predict part performance, which in turn can support accept or reject decisions for printed parts. This paper proposes a framework for incorporating ISM-generated information, with a particular focus on infrared (IR) image analysis for this purpose. IR camera images, in conjunction with numerical features such as infill pattern and extruder nozzle temperature, serve as an input to a multimodal deep learning (MDL) model that predicts the mechanical performance of printed parts. In the framework, convolutional neural nets process image inputs, while a
Mollan, Calahan, Kulkarni, Saurabh, Malik, Ali Ahmad, Patterson, Albert E., Pandey, Vijitashwa
Machine Learning for Predicting Rotor and Brake Fluid Temperatures in Hill Descent Events2026-01-0159To be published on 04/07/2026
High thermal loads on brake systems during extended descents followed by vehicle soak pose significant safety and durability risks. Excessive rotor or fluid temperatures can cause loss of braking efficacy, fluid degradation or evaporation, thermal fade, and accelerated component wear. This study uses time history data of brake disc and fluid temperatures collected during controlled hill descent events and subsequent soak (park) periods, and simultaneously logs environmental conditions (ambient temperature, humidity, wind speed and direction) and brake specifications (rotor, caliper geometry, pad material, fluid type, and vehicle mass). Seventy six test runs from a dedicated validation program are used, each providing time history records that form the basis of our analysis. From these records we extract phase specific samples (descent and soak phase) and engineer compact descriptors, such as start and peak temperatures, environmental factors, rolling statistics and contextual metadata
poojari, Uday Kumar, Westphalen, Jan, Venugopal, Narayana
Curvature-Based End-to-End Autonomous Vehicle Path Planning at Intersections2026-01-0045To be published on 04/07/2026
Autonomous vehicle navigation requires accurate prediction of driving path curvature to ensure smooth and safe trajectory planning. This paper presents an end-to-end approach to curvature prediction using deep neural networks trained on real-world driving data. We develop a comprehensive neural network architecture that directly maps high-dimensional sensor inputs to curvature predictions, eliminating the need for intermediate feature engineering steps. The model processes multi-modal inputs including vision features, vehicle state parameters, and environmental context through a deep learning pipeline. Our end-to-end approach demonstrates the feasibility of learning complex driving behaviors directly from raw sensor data, providing a more robust and generalization solution compared to traditional rule-based methods. The neural network architecture incorporates dropout regularization and adaptive learning rate scheduling to ensure stable training and optimal performance. Results show
Hajnorouzali, Yasaman, Wang, Hanchen, Li, Taozhe, Burch, Collin, Lee, Victoria, Tan, Lin, Arjmandzadeh, Ziba, Xu, Bin
Roundabout Dilemma Zone Data Mining and Forecasting with Trajectory Prediction and Graph Neural Networks2026-01-0029To be published on 04/07/2026
Traffic roundabouts, as complex and safety-critical road scenarios, present significant challenges for autonomous vehicles. In particular, predicting and managing dilemma zone (DZ) encounters at roundabout intersections remains a pivotal concern. This paper introduces an AI-driven system that leverages advanced trajectory forecasting to anticipate DZ events, specifically within traffic roundabouts. At the core of our framework is a modular, graph-structured recurrent architecture powered by graph neural networks (GNNs). By modeling agent interactions as a dynamic graph, our approach integrates heterogeneous data sources, including semantic maps, while capturing agent dynamics with high fidelity. This GNN-based forecasting model enables accurate prediction of DZ events and supports safer, data-driven traffic management decisions for both autonomous and human-driven vehicles. We validate our system on a real-world dataset of roundabout intersections, where it achieves high precision with
Lu, Duo, Farhadi, Mohammad, Satish, Manthan, Yang, Yezhou, Chakravarthi, Bharatesh
Speed Determination Using Audio of Dash Camera Video from Commercial Vehicle Tires Frequencies for Vehicle Accident Reconstruction2026-01-0536To be published on 04/07/2026
Prior research has validated a reliable method for determining vehicle speed using audio recorded by cameras mounted in vehicles, specifically for rolling passenger vehicle tires. Passenger vehicle tires produce a frequency component directly correlated to vehicle speed when traveling on concrete roadways. However, prior research has not been conducted on audio for rolling commercial vehicle tires, which differ in construction from passenger vehicle tires and therefore produce audio signals on roadway surfaces that passenger vehicles tires did not when tested in the prior study. The current research concluded that commercial vehicle tires rolling on various roadway surfaces also generated a frequency that varied with vehicle speed. The purpose of this study was to outline, test, and confirm the source of the speed-dependent frequency and to develop a validated method for use in forensic applications. A modified version of the passenger vehicle tire equation from prior research was
Vega, Henry V., Cornetto, Anthony, Ngo, Long Justin, Hatab, Ziad, Hunter, Eric
Thermal Management Simulation of Liquid-Cooled Energy Storage Batteries Using a Reduced-Order Model2026-01-0122To be published on 04/07/2026
A linear time-invariant (LTI) reduced-order model (ROM) based on 3D CFD offers high accuracy for simulating energy storage battery thermal behavior under variations in ambient temperature, initial temperature, coolant inlet temperature, cell heat generation, and Joule heating. However, it cannot capture coolant flow rate fluctuations, which are essential for liquid cooling start–stop conditions in energy storage containers. To address this, we use an enhanced LTI+HTC ROM, which incorporates flow-rate-dependent heat flux at the fluid–solid interface into the LTI ROM. This is achieved by establishing a correlation between the coolant flow rate and the interface heat transfer coefficient (HTC). We further improve model fidelity by applying interface heat flux distribution features extracted from CFD. Comparisons with 3D CFD results demonstrate that the LTI+HTC ROM is highly accurate and significantly reduces computational cost, making it a practical tool for thermal management studies
guo, Jia, CHEN, Guijie, ma, Shihu, Hu, Xiao, Jing, Li, SONG, Shujun, Huang, Long
Autonomous Racing Control Using Reinforcement Learning with Racing Line Guidance2026-01-0031To be published on 04/07/2026
In recent years, with the emergence of autonomous racing competitions such as Roborace, autonomous racing cars have become a prominent research focus and have undergone rapid development. This paper establishes a reinforcement learning-based decision-making and control framework for autonomous racing cars, achieving cooperative optimization between high-speed racing and safety control under extreme operating conditions. By integrating a Control Barrier Function constrained reward function for dynamic stability and a racing line-guided path optimization, the proposed framework achieves a synergistic balance between minimum lap time and operational safety. Dynamic stability domain constraints based on a two-degree-of-freedom single-mass rigid-body model and racing line optimization based on track geometric modeling are established. The phase plane stability boundary is constructed using the yaw rate method and double-line method, upon which Control Barrier Function constraints are
peng, tao, Zhang, Weiqi, Li, Zengwen, sudong, jiang
Prediction of Battery Self-Discharge Under Sparse Data Using Physics-Guided Machine Learning2026-01-0155To be published on 04/07/2026
Predicting battery self-discharge across wide temperature ranges and extended durations remains a significant challenge due to the scarcity of physical test data, which is typically limited to a few temperature points and short observation windows. This limitation complicates generalization and increases the risk of inaccurate extrapolation. To address this, the paper introduces a machine learning–based framework designed to predict self-discharge behavior under diverse thermal conditions and long time horizons. Multiple modeling strategies are examined, including feedforward neural networks, long short-term memory (LSTM) architectures, synthetic data generation, and physics-informed integration of governing equations. Particular emphasis is placed on hybrid and physics-regularized models that embed first-principles relationships to guide extrapolation beyond the observed data domain. This approach mitigates the inherent instability and potential errors associated with purely data
Chavare, Sudeep, Zeng, Yangbing, Muppana, Sai Siddhartha, Miao, Yong, Xu, Simon
Leveraging Machine Learning for Occupant Body Size Estimation via B-Pillar Seatbelt Systems2026-01-0559To be published on 04/07/2026
Occupant body size in vehicles varies significantly, encompassing differences in height, weight, and overall body composition. Adaptive restraint systems, featuring adjustable parameters such as belt load limiters, steering column load limiters and stroke, seat pan stiffness, and airbag pressure, can offer more equitable protection tailored to individual body sizes. In this study, a test rig modeled after the Volvo XC90 was used to collect data from 47 seated participants. Key seatbelt-related parameters, including D-ring angle, belt payout length, lap belt length, and buckle tension, were measured. These measurements were used to train machine learning models to predict occupant characteristics: height, weight, sitting height. The results demonstrate that low-cost sensors embedded in the seatbelt system can provide sufficiently accurate occupant body size estimations to inform adaptive restraint systems. The prediction errors across both training and test datasets were as follows
Wang, Da, Ahmed, Jawwad, Rowe, Mike, Brase, Dan
Bias and uncertainty in the time, position, and speed data of consumer-grade GPS devices2026-01-0544To be published on 04/07/2026
The goal of this study is to quantify the accuracy and uncertainty of position and speed measurements acquired by a wide range of consumer-grade GPS-enabled devices while road cycling. We acquired position and speed data simultaneously from 13 consumer devices (e.g., phones, watches, bicycle computers) at their maximum sampling rate (typically 1 Hz) during 10 hours (thirty 20 minute sessions) and 150 km of road cycling in a suburban environment. The position data from these devices were compared to real-time kinematic (RTK) data acquired using a differential GPS system and the speed data from these devices were compared to speed data acquired with high-resolution wheel rotation sensors (23,040 pulses per revolution) synchronized with the RTK data. Based on these comparisons, we generated probability distributions for the errors in the position and speed measured by each of the consumer devices. This study provides foundational data needed to quantify the accuracy and uncertainty across
Booth, Gabrielle R., Mitchell, Alan L., Siegmund, Gunter P.
Determining Video Frame Capture Rate through Reverse Engineering2026-01-0543To be published on 04/07/2026
The timing of video recordings, along with the spatial positioning of objects, is a fundamental parameter for calculating the speed time history. If the task involves determining an object’s average speed over several to a dozen frames, the average inter-frame interval can be used as the basic time unit. However, if the objective is to compute speed from individual frames, the reliability of the timing becomes crucial. Without access to DVR hardware documentation, proprietary algorithms, or software – and considering the frequent hardware modifications and software updates - the most effective way to solve the problem is through a reverse-engineering approach. This study discusses several aspects of timing analysis, including: (1) making a test recording of a calibrated LED lightboard; (2) analyzing the relationship between the lightboard time and the presentation time stamp (pts) extracted from the file metadata; (3) investigating frequency interference occurring when recordings are
Wach, Wojciech
From Prediction to Design: Using Context-Aware Graph Neural Networks and Explainable AI to Anticipate Transfer-of-Control2026-01-0049To be published on 04/07/2026
Drivers often interact with partial automation (SAE Level 2) systems, initiating transfer of control (TOC) either by handing control over to the automation or by taking it back. Accurately predicting these interactions may inform the design of future automation that adapts proactively to the operating context, enhances comfort, and improves safety. We present a context-aware framework that generates a unified driver–vehicle–environment representation by fusing data from videos of the in-cabin and forward road with vehicle kinematics, and driver behaviors (annotations of driver glance, hands-on-wheel, and secondary-task engagement). This representation is encoded in a hierarchical Graph Neural Network that classifies driver-initiated TOC to: (i) Manual-to-Automation (M2A) and (ii) Automation-to-Manual (A2M) and predicts time-to-TOC. Shapley-based explainable AI was used to quantify how the importance of behavioral, contextual, and kinematic cues evolved in the seconds preceding a TOC
Zhao, Zhouqiao, Gershon, Pnina
The cure for overgrown process garden or why engineers have given up with ASPICE.2026-01-0067To be published on 04/07/2026
The useability of development processes in automotive decreased in the past years to a level at which their application and being able to benefit from them are questionable. Such degradation can be attributed to new additions to the processes and introduction of FUSA and Cybersecurity standards. The processes try to keep up with the shift from “plan-implement- test-roll out” to more agile methods. In addition, process departments, in charge of these processes, did not target the engineers as users of these processes, they focused on the compliance of processes and the assessor in the assessment. Due to reasons mentioned above, processes have grown in size and complexity which creates unnecessary overhead and forces the usage of complex tools to handle this overhead. Furthermore, the language of this process is written mostly in the language of the assessor, making them hard to understand for an average engineer. As a result, engineers are annoyed in their daily work and are not as
Weber, Matthias, Kmiec, Mateusz, Romijn, Marcel, Nedkov, Detelin
Method of Estimating Electric Vehicle Charge Time Based on Neural Networks2026-01-0382To be published on 04/07/2026
Accurate prediction of electric vehicle charging time is critically hindered by dynamic, non-linear factors including battery aging which is indicated by the State of Health (SOH), substantial power diversion to thermal management systems in extreme temperatures, fluctuating user-defined accessory loads, and hardware limitations of the charging infrastructure. Traditional estimation methods, reliant on static models or predefined calibrations, fail to adapt to these real-world variables, leading to inaccurate predictions and user dissatisfaction. This paper presents a novel data-driven estimation framework utilizing a tailored feedforward neural network architecture specifically designed for this complex task. The model processes a sensitive set of inputs—including initial State of Charge (SOC), SOH, battery temperature, charging station power level and user-selected target SOC—to effectively capture the intricate, non-linear interdependencies governing the charging process. The
Xie, Zhentao, shojaei, Sina, Weslati, Feisel
Attention and Intrinsic Curiosity-Enhanced Deep Reinforcement Learning for Path Planning in Dynamic Environments2026-01-0044To be published on 04/07/2026
Efficient and reliable path planning remains a core challenge for autonomous vehicles operating in dynamic and crowded environments. Although Deep Reinforcement Learning (DRL) has shown considerable potential in autonomous decision-making, it still faces challenges such as insufficient feature extraction, sparse rewards, and low obstacle avoidance efficiency in complex scenarios. To address these issues, this paper proposes an end-to-end path planning framework, PPO-ICM-Attn. Built upon the Proximal Policy Optimization (PPO) algorithm, the framework incorporates a dual-channel attention convolutional neural network module (Attention-CNN) to enhance spatial and semantic understanding of dynamic obstacles, and introduces an Intrinsic Curiosity Module (ICM) to promote active exploration in sparse reward settings. Furthermore, a reactive avoidance reward function based on velocity obstacle theory is designed and embedded to achieve real-time proactive collision avoidance in highly dynamic
Shen, Shiquan, Liu, Jiahao, Chen, Zheng, Li, Zongdian, Zhao, Yuting, Wu, Minggong, Zhao, Jie, Qin, Zongquan, Wang, Yanfeng
Techno-Economic Evaluation of Electified Vehicle Options in Drayage Fleets2026-01-0454To be published on 04/07/2026
The transition of drayage fleets from diesel to battery electric vehicles (BEVs) poses both opportunities and uncertainties, especially regarding long-term economic viability. While total cost of ownership (TCO) is a useful benchmark, fleet operators and investors are equally concerned with investment performance metrics such as payback period and Internal Rate of Return (IRR), which better reflect financial risks and investment return timelines. This study proposes a comprehensive framework to evaluate the economic feasibility of transitioning from diesel to BEVs, jointly analyzing TCO parity conditions, payback periods, and IRRs under different technology, cost, and operational scenarios. Building on a previously developed cost modeling tool, the analysis incorporates evolving battery prices, energy costs, vehicle performance characteristics and other assumptions from 2025 to 2050. Real-world duty cycle data from a Class 8 drayage fleet at the Port of Savannah is used to capture
Sun, Ruixiao, Sujan, Vivek, Goulet, Nathan, Wang, Qixing
Data-driven Study on Chassis Suspension Performance Degradation2026-01-0582To be published on 04/07/2026
The performance of chassis suspension mechanisms critically affects vehicle handling, ride comfort, and safety. Under severe operating conditions, health monitoring of chassis suspension mechanisms can detect potential problems in time, thereby reducing safety risks and optimizing maintenance costs. With the rapid development of sensing and data analytics technologies, data-driven approaches are increasingly used in health monitoring. This study aims to achieve real-time monitoring of chassis suspension performance degradation using a deep neural network (DNN). First, a semi-vehicle suspension model is constructed to simulate the vehicle's degraded state under bumpy road conditions, generating a dataset of key suspension parameters. Subsequently, a DNN model comprising three hidden layers is developed to assess suspension performance degradation. To optimize model performance, the effects of different numbers of neurons and hidden layers on model accuracy are explored. Experimental
Liao, Yinsheng, Lei, Yisong, Su, Ailin, Wang, Zhenfeng
Securing the AI-Driven Vehicle: A Hardware-Based Security Framework for Scalable, Trustworthy Autonomy2026-01-0085To be published on 04/07/2026
Autonomous driving technology in modern vehicles marks a pivotal evolution in transportation, but it may also introduce system-level vulnerabilities that span from tampering with sensors and interfaces to compromising compute units and communication links. This paper proposes a unified, layered hardware security architecture for AI-equipped automated vehicles. Based on current automotive Ethernet and zonal architectures, it provides end-to-end trust using hardware interface security, accelerated-cryptography, and SRAM PUF-based key provisioning. All security primitives are anchored to a hardware root of trust, delivering cryptographic identity, secure boot enforcement, and trusted key storage across the entire vehicle lifecycle. Our architectural assumptions follow the direction ISO 26262, ISO/SAE 21434 and AI development mandates for next-generation SAE Level 3+ vehicles using zonal Ethernet backbones, Time-Sensitive Networking (TSN), and heterogeneous compute islands (e.g., vision
C Suriyanarayanan, Pav, Iacob, Radu
Automotive Crash Dynamics Modeling Accelerated with Machine Learning2026-01-0568To be published on 04/07/2026
Crashworthiness assessment is a critical aspect of automotive design, traditionally relying on high-fidelity finite element (FE) simulations that are computationally expensive and time-consuming. This work presents an exploratory comparative study on developing machine learning-based surrogate models for efficient prediction of structural deformation in crash scenarios using the NVIDIA PhysicsNeMo framework. Given the limited prior work applying machine learning to structural crash dynamics, the primary contribution lies in demonstrating the feasibility and engineering utility of the various modeling approaches explored in this work. We investigate two state-of-the-art neural network architectures for modeling crash dynamics: MeshGraphNet, a graph neural network that is widely employed in physics-based simulations, and Transolver, a transformer-based architecture with a physics-aware attention mechanism designed to maintain linear computational complexity with respect to geometric
Nabian, Mohammad Amin, Chavare, Sudeep, Akhare, Deepak, Ranade, Rishikesh, Cherukuri, Ram, Tadepalli, Srinivas
From CFD Simulations to Machine Learning: A Comparative Study of ML Tools for Aerodynamic Drag Prediction2026-01-0598To be published on 04/07/2026
This study presents the development of machine learning frameworks for predicting aerodynamic drag coefficients in automotive vehicle designs using data derived from computational fluid dynamics (CFD) simulations. The input data includes both structured tabulated design parameters and unstructured geometric representations in the form of point clouds, enabling a multi-modal approach to aerodynamic modeling. A range of machine learning models was explored, from traditional regression techniques to advanced deep learning architectures such as graph neural networks (GNNs), each tailored to handle specific data formats. These frameworks were designed to integrate seamlessly with CFD workflows, supporting efficient training and evaluation using standard regression metrics. The work aims to establish a foundation for data-driven aerodynamic analysis in automotive design, emphasizing the role of machine learning in enhancing simulation-based engineering processes.
Kumar, Gaurav, Khanna, Susheel
An Updated Correlation Study between the Full Size Wind Tunnels of General Motors, Ford Motor Company, and FCA US LLC2026-01-0612To be published on 04/07/2026
As automotive aerodynamic testing facilities evolve to capture more real-world behavior, updating the correlation between old and new technologies is essential. Recently, the three-member consortium of the United States Council for Automotive Research (USCAR) - FCA US LLC, Ford Motor Company, and General Motors - transitioned from full-size static ground plane facilities to 5-belt moving ground plane wind tunnel facilities. The primary objective of this study was to update the correlation data sets to maintain consistent and robust data sharing among companies, which is the cornerstone of USCAR efforts. To achieve this, a set of updated correlation data sets were calculated to replace the original correlation study results from 2008. Additionally, the methodology for applying correlation equations was revised from using averaged wind tunnel data to employing direct wind tunnel-to-wind tunnel correlation equations. In a two-phase correlation effort conducted in 2022 and 2025, the three
Nastov, Alexander, Lounsberry, Todd, Madin, Trevor, Langmeyer, Gregory, Fadler, Gregory, Skinner, Shaun, Horton, Damien
Adaptive Torque Vector Control for Distributed Drive Intelligent Electric Vehicles Using Model and Data Driven Approaches2026-01-0628To be published on 04/07/2026
To effectively improve the performance of chassis control of distributed drive intelligent electric vehicles (EVs) under difference road conditions, especially in combing road information and chassis control for improving road handling and ride comfort, is a challenging task for the distributed drive intelligent EVs. Simultaneously, inaccurate chassis control and uncertainty with system input, are always existing, e.g., varying road input or control parameters. Due to the higher fatality rate caused by variable factors, how to precisely chose and enforce the reasonable chassis control strategy of distributed drive intelligent EVs become a hot topic in both academia and industry. To issue the above mentioned, an adaptive torque vector hierarchical controller based on road roughness and adhesion is proposed, which optimizes the comprehensive performance of vehicle chassis. First, combined with the characteristic of the unbalance dynamic force caused by the air gap between the stator and
Wang, Zhenfeng, Zhao, Gaoming, Zhang, Zhijie
Optimization of a Feedforward Neural Network Machine Learning Battery Model for an Automotive Pouch Cell2026-01-0386To be published on 04/07/2026
Modern battery management systems have a critical need for highly accurate battery terminal voltage models, which are a key component of algorithms which estimate or predict power capability, range, temperature, and other factors. While electro-chemical and equivalent circuit models are widely used for this purpose, they typically struggle to capture the complex, non-linear dynamics and degradation effects inherent in real-world battery operation, necessitating frequent recalibration. This study addresses these limitations by proposing a robust, data-driven approach for terminal voltage estimation using a feedforward neural network (FNN) machine learning model. An LG E66 cell, a high-performance lithium-ion battery with a Nickel Cobalt Manganese (NCM) cathode, was selected for this study due to its prevalent use in light duty electric vehicles. Data collected at temperatures ranging from 40 °C to -20 °C is used to train and test the models, with model error reported for HWFET, UDDS
Dehury, Biswanath, Nahidmobarakeh, Lucas, Panchal, Satyam, Gross, Oliver, Kollmeyer, Phillip
Machine Learning Modeling for Compressive Property Prediction of Foams2026-01-0512To be published on 04/07/2026
Foam material models for automotive structural analysis typically require tensile and compressive data at multiple strain rates. The testing is costly and may require a long time to complete. For many applications, foams of similar chemistry are used and the foam structural responses, such as stiffness and compression force deflection, are controlled by the foam density. In such cases, Machine Learning (ML) lends itself as an ideal tool to detect the trends in material response based on density and strain rate. In this paper, five microcellular polyurethane (PU) foams of different densities were tested at four strain rates ranging from 0.01/s to 100/s. A ML model capable of predicting compressive stress-strain response for a range of densities was developed. The model demonstrated good prediction capability for intermediate strain rates at all foam densities and in extrapolating stress-strain curves at higher densities at all strain rates. The strain rate trends for a density outside
M, Gokula Krishnan, Kavimani, Harish, Muppana, Sai Siddhartha, Savic, Vesna, Chavare, Sudeep, V S, Rajamanickam
Design and Evaluation of a Deep Neural Network Model for Prediction of Diesel Equilibrium Combustion Products and Characteristics2026-01-0293To be published on 04/07/2026
Accurate prediction of equilibrium combustion products and thermodynamic properties is essential for optimizing engine performance, enhancing combustion efficiency, and reducing emissions in diesel-powered systems. Traditional methods for combustion modeling often involve solving complex chemical equilibrium equations or thermodynamic relations, which could be computationally expensive and time-consuming. In this study, we present a data-driven approach using a deep neural network (DNN) model to predict the equilibrium combustion products and key thermodynamic characteristics of diesel under varying thermodynamic conditions. The proposed DNN model is trained on a comprehensive dataset generated from equilibrium calculations. The inputs include pressure, temperature, and equivalence ratio, covering a relatively wide range to encompass diesel equilibrium combustion under various conditions. Outputs are equilibrium combustion products and thermodynamic properties, including enthalpy
Ji, Huangchang, Wang, Kai, Guo, Zhefeng, Han, Yang, Lee, Timothy
Ambient and Initial Temperature Effects on Battery Electric Vehicle (BEV) Range and Energy Consumption Rate Modeled in FASTSim2026-01-0277To be published on 04/07/2026
Ambient and initial temperatures have a significant impact on the energy consumption rate of battery electric vehicles (BEVs) due to auxiliary loads and the dependence of battery efficiency on temperature. Quantifying these impacts can help BEV owners, communities, and researchers understand and prepare for the impact of temperature on BEV vehicles, especially as adoption increases. We used NREL's FASTSim 3 to investigate the impacts of a two-dimensional design-of-experiment that swept ambient and initial temperature in simulating a compact BEV hatchback over UDDS (city) and HWFET (highway) US EPA drive cycles. Compared to a 22 ° C ("cozy") initial / ambient temperature baseline in UDDS, a cold initial / cold ambient temperature (-7 ° C) resulted in a 221% increase in the energy consumption rate, a cozy initial / cold ambient temperature resulted in a 94% increase, a hot initial (45 ° C) / hot ambient (40 ° C) temperature resulted in a 40% increase, and a cozy initial / hot ambient
Baker, Chad, Steuteville, Robin, Holden, Jake, Gonder, Jeffrey, Carow, Kyle
Items per page:
1 – 50 of 6184