Browse Topic: Artificial intelligence (AI)

Items (1,396)
Find
testsofware-test-123jms4/1/2026
test
SAE TOMORROW TODAY - Building an End-to-End Autonomous Driving System135572/27/2026
Join us as we sit down with Richard Chelminski, Senior Vice President, Head of SDV Platform, at 42dot, an autonomous driving software and mobility platform development startup from Hyundai Motor Group. We discuss how the company is pushing boundaries with its end-to-end autonomous driving system, Atria AI, which is focused on Level 4 autonomy and designed for seamless integration into consumer vehicles -- no LIDAR required! You'll also learn why camera-first and radar technologies are reshaping the industry, how software is now at the heart of vehicle development, and what the future holds for Transportation as a Service (TaaS). If you're curious about the next wave of smart, software-defined vehicles and autonomous driving innovation, this episode is a must-listen. We'd love to hear from you. Share your comments, questions and ideas for future topics and guests to podcast@sae.org. Don't forget to take a moment to follow SAE Tomorrow Today--a podcast where we discuss emerging
Patterson, Lori
Artificial Intelligence in Passenger Cars: Unlocking Intelligent Mobility and Human-Centric Vehicle Innovation2026-26-06721/16/2026
Artificial Intelligence (AI) is fundamentally reshaping the landscape of automotive engineering, particularly in passenger vehicles where comfort, safety, and customization are paramount. This paper presents a comprehensive review of current and emerging AI-driven applications in passenger cars, including intelligent vehicle insurance, real-time traffic optimization, personalized driving experiences, emotion recognition, and advanced EV battery management. The integration of AI technologies—ranging from machine learning and computer vision to natural language processing and deep neural networks—enables predictive maintenance, adaptive driver-assist systems, and next-generation human-machine interfaces. AI’s role in vehicle insurance is also explored through behaviour-based risk scoring and claims automation. Furthermore, emotion-aware cabin systems are discussed, which aim to detect driver fatigue, stress, or distraction to improve road safety and mental wellness. The paper also
Hazra, Sandip, Tangadpalliwar, Sonali, Khan, Arkadip
Enhancing Engine Reliability Testing with IoT, AI and ML: Real-Time ML-based Limit Monitoring and Predictive Maintenance with IoT sensor2026-26-06471/16/2026
The integration of Internet of Things (IoT), Artificial Intelligence (AI), and Machine Learning (ML) has transformed various industries, offering substantial benefits. The application of these technologies in engine reliability testing has immense potential as they offer real-time monitoring and analysis of engine performance parameters. Engine reliability testing is vital for ensuring the safety, efficiency, and longevity of engines. Traditional methods are time consuming, expensive, and rely heavily on manual inspection and data analysis. This paper shows how IoT, AI and ML technologies can greatly enhance the efficiency of engine reliability testing. The paper includes the following case studies: 1. ML-based Limit Monitoring for test data: With the help of machine learning models, we monitor the Engine health by comparing predicted and measured values during real-time testing. Intelligent algorithms predict and compare values, detect anomalies, and identify root causes in real-time
Yadav, Sanjay Kumar, kumar, prabhakar, R, Dinesh, Joon, Sushant, Rai, Ayush, Tripathi, Vinay Mani
Next-Gen Driver Health and Wellness Platform for Fatigue Detection and Emotional Wellbeing Monitoring2026-26-06391/16/2026
Driver fatigue and emotional distress are major contributors to traffic accidents, especially among long-haul and professional drivers, making their detection a critical area in road safety and public health. While existing research has explored various sensor-based and wearable solutions to monitor driver states, these approaches often suffer from high cost, intrusiveness, or limited scalability. Camera-based systems have recently gained attention for being less invasive and more practical for real-world deployment; however, many still depend on multiple sensors or cloud-based processing, which raise concerns around energy consumption, privacy, and integration complexity. To address these gaps, we present a sustainable, camera-only Driver Health and Wellness Platform for real-time fatigue detection and emotional wellbeing monitoring. The system leverages a single in-cabin camera and applies lightweight, edge-optimized computer vision models to analyze facial features, eye dynamics
Iqbal, Shoaib, Imteyaz, Shahma
A Large Concept Model Approach for Summarizing Automotive Documents and Enhancing Vehicle Data Query2026-26-06761/16/2026
Technological advancement in automobile production, regulations, sales, and services leads to extensive technical documents and manuals. However, lengthy documentation, language barriers make it challenging to find quickly, highlighting the importance of summarization to bridge this knowledge gap. The Large Language Models (LLMs) have demonstrated impressive capabilities in tasks like Q&A, coding, drafting, and scientific acumen, and are commonly used to condense many complex documents. Despite LLMs being widely used in various sectors like research and journalism, LLMs face challenges in reliably summarizing content. Their word-by-word tokenization can sometimes fail to capture the true meaning and context, especially for abstractive summaries. This paper explores Large Concept Models (LCMs), which operate on principles of semantic reasoning, cross-modality integration, and hierarchical structuring at a higher conceptual level. This makes LCMs language and modality agnostic
Singh, Samagra, Ravi, Utkarsh, Vikram, Prateek, Shenoy, Lakshmi, Awasthi PhD, Anshuman
Machine Learning Approach to ISO Road Class Prediction Using Multiple Linear Regression2026-26-06671/16/2026
The inertial profiler methodology is traditionally employed in RLDA (Road Load Data Acquisition) to measure road profiles and classify test routes into ISO road classes. However, this approach demands significant time and effort during instrumentation. Also, during data acquisition, laser height sensor data is affected especially during adverse conditions such as rainy seasons or on surfaces with improper reflectivity. Additionally, substantial resources are required for data processing to convert raw measurements into road classifications. To address these challenges, an initial attempt was made to establish a relationship between axle acceleration responses and road profiles, enabling axle acceleration measurements during RLDA to predict ISO road classes. However, this approach relied on a simple linear model that considered only axle acceleration responses, rendering the predictions susceptible to inaccuracies due to varying parameters such as vehicle speed. To overcome these
P, Praveen Kumar, P, Dayalan, Sriramulu, Yoganandam
Unlocking Edge AI in Software-Defined Vehicles: Balancing model complexity, compute resources, and data access2026-26-06861/16/2026
The potential of Edge AI in automotive goes beyond ADAS and automated driving. AI-driven optimizations in driving efficiency, in-cabin personalization, predictive maintenance, and enhanced safety require intelligent processing at the vehicle edge. However, integrating AI into current vehicle electrical/electronic architectures (EEAs) presents significant challenges. This paper examines how AI models can run on general-purpose ECUs, gateways, and domain controllers while balancing model complexity, compute resources, and data availability. We explore trade-offs, architectural choices, and the need for software configurability to ensure AI models evolve with new data and vehicle parameters. We conclude by providing a glimpse into Sonatus’ latest AI product that encapsulates a holistic approach optimized for vehicle platforms—spanning cloud-based model development, optimization, and validation to in-vehicle inference—that unlocks the potential of Edge AI and SDVs for the automotive
Sah, Mohamadali, Khatri, Sanjay
Emerging Trends in Emission Optimization2026-26-02291/16/2026
The exhaust emission control is a vital aspect of automotive development, aimed at ensuring effective control of pollutants such as NOx, CO, and HC. The traditional method of calibrating emission control strategies, that requires extensive vehicle testing with both the fresh and aged catalysts under a variety of operating conditions, is a time-consuming process. The frequent changes in emission legislation creates new challenges for achieving a faster time to market and effective utilization of resources. So, the use of automation and machine learning (ML) in the domain of emission control is the need of the hour to proactively tackle these challenges. This paper attempts to explore emerging trends in emission optimization, such as the automated measurement process and ML/AI based data analysis, that can improve the overall process efficiency in the calibration of emission control strategies. The automated measurement process is achieved using the automation software. The data
Dhayanidhi, Hukumdeen, Balasubramanian, Karthick, A, Akash
Transforming Gear Manufacturing_ A Deep Learning Approach to Quality Prediction2026-26-06461/16/2026
The automotive industry is rapidly transitioning towards Industry 4.0, transforming vehicle manufacturing. To achieve a lower carbon footprint, it is crucial to minimize raw material wastage and energy consumption. Reducing component wastage, lead time, and automating gear manufacturing are key areas. Gear micro-geometry inspection is vital, as variations affect service life and NVH (Noise, Vibration, Harshness). Despite standards for permissible errors, manual evaluation is often needed.This subjective evaluation approach will have a possibility that a gear with undesired variations gets assembled into the product. These issues can be detected during NVH testing, leading to replacement of part and re-assembly thus increasing lead time. This generates a need for an automated system which could reduce the human intervention and perform the activity. This paper discusses methodologies to meet these requirements and achieve desired results. The research aims to develop a deep learning
Ramakrishnan, Gowtham Raj, Baheti, Palash, PR, Vaidyanathan, Durgude, Ranjit, Bathla, Archana, R, Greeshmita, V, Rangarajan
Real-Time In-Cabin Vital Sensing and Emergency Response Using Ultra-Wideband Radar in Automotive Environments2026-26-00121/16/2026
This study introduces a novel in-cabin health monitoring system leveraging Ultra-Wideband (UWB) radar technology for real-time, contactless detection of occupants' vital signs within automotive environments. By capturing micro-movements associated with cardiac and respiratory activities, the system enables continuous monitoring without physical contact, addressing the need for unobtrusive vehicle health assessment. The system architecture integrates edge computing capabilities within the vehicle's head unit, facilitating immediate data processing and reducing latency. Processed data is securely transmitted via HTTPS to a cloud-based backend through an API Gateway, which orchestrates data validation and routing to a machine learning pipeline. This pipeline employs supervised classifiers—Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Random Forest (RF)—to analyze features such as temporal heartbeat variability, respiration rate stability, and energy distribution patterns
Singh, Samagra, Pandya, Kavita, Jituri, Keerti
Artificial Neural Network for predicting thermal cool-down behavior of a School Bus Cabin2026-26-06531/16/2026
Passenger vehicles, such as buses, exhibit significant heat absorption when parked under direct sunlight, leading to elevated interior temperatures during daytime and a consequent reduction in thermal comfort. In our previous study, we developed and validated a transient numerical method to predict the thermal behavior of the cool-down process in a school bus cabin. Despite its accuracy, modeling such complex simulations using Computational Fluid Dynamics (CFD) remains time-intensive, with pre-processing taking up to a week and simulations requiring an additional 8-10 hours. While physics-based simulations are known for their accuracy, they are inherently slower. Conversely, data-based meta models offer faster results, but their outputs must be carefully correlated with test data or physics-based model data. This study aims to leverage existing CFD data to train and develop an Artificial Neural Network (ANN) based meta model, which is then benchmarked against test results for
Suryavanshi, Apoorva, sharma, shantanu
Reverse Engineering and Digital Twins in Powertrain Design: Enhancing Performance and Sustainability2026-26-00651/16/2026
This study explores the application of reverse engineering (RE) and digital twin (DT) technology in the design and optimization of advanced powertrain systems. Traditional approaches to powertrain development often rely on legacy designs with limited adaptability to modern efficiency and emission standards. In this work, we present a methodology combining 3D scanning, computational modeling, and machine learning to reconstruct, analyze, and enhance internal combustion engines (ICEs) and electric vehicle (EV) drivetrains. By digitizing physical components through RE, we generate high-fidelity DT models that enable virtual testing, performance prediction, and iterative improvement without costly physical prototyping. Key innovations include a novel mesh refinement technique for scanned geometries and a hybrid simulation framework integrating finite element analysis (FEA) and multi-body dynamics (MBD). Our case study demonstrates a 12% increase in thermal efficiency for a retrofitted ICE
Bernikov, Mark Alexandrovich, Kurmaev, Rinat
Lightweight AI Deployment for Legacy Automotive ECUs: A Practical Optimization Approach2026-26-06631/16/2026
Automotive systems are increasing adopting data-driven and intelligent functionality in the areas of predictive maintenance, virtual sensors and diagnostics. This has led to a need for the AI models to be directly run on vehicle ECUs. However, most of these ECUs – especially those in cost-sensitive or legacy platforms lack the computational capacity and parallel processing support required for standard AI implementations. Given the stringent real-time and reliability requirements in automotive environments, deploying such models presents a unique challenge. This paper proposes a practical methodology to optimize both the training and deployment phases of AI models for low-computation ECUs that operate without parallelism. Designing lightweight model architectures, using pruning and quantization techniques to minimize resource utilization, and putting in place a strategy appropriate for single-threaded execution are the three main objectives of the developed approach. The goal is to
Sharma, Sahil, Mathew, Melvin
NN FMU: Deep Learning Models for Next-Gen xiL Vehicle Simulation2026-26-04521/16/2026
Functional Mock-up Units (FMUs) have become a standard for enabling co-simulation and model exchange in vehicle system development. However, traditional FMUs derived from physics-based models can be computationally intensive, especially in scenarios requiring real-time performance. This paper presents a Python-based approach for developing a Neural Network (NN) based FMU using deep learning techniques, aimed at accelerating vehicle simulation while ensuring high fidelity. The neural network was trained on vehicle simulation data and developed using Python frameworks such as TensorFlow. The trained model was then exported into FMU, enabling seamless integration with FMI-compliant platforms. The NN FMU replicates thermal behavior of vehicle with high accuracy while offering a significant reduction in computational load. Benchmark comparisons with Physical Thermal Model demonstrate that the proposed solution provides both efficiency and reliability across various driving conditions
Srinivasan, Rangarajan, ASHOK BHARDE, Pooja, MHETRAS, Mayur, CHEHIRE, Marc
Accelerating engine auxiliaries’ development by deploying ML techniques in CFD virtual verification2026-26-03741/16/2026
The automotive and off-road industries are heavily investing in research and development (R&D) to improve both physical and virtual verification and validation techniques. Recent software and hardware advancements have extended these techniques from simple component evaluations to complex system assessments such as involving multi-physics scenarios. Despite the benefits of virtual validation tools like structural analysis and computational fluid dynamics (CFD), they often come with high development costs, particularly in CFD applications. Virtual verification methodology, especially when combined with data science, offer significant advantages over traditional physical methods by enhancing computer-aided engineering (CAE) efficiency and reducing resource consumption which can greatly improve product design and validation efficiency across many industries. The success of machine learning applications depends on effective data processing, adequate computational resources, and the right
Jadhav, Mitali, Kumbhar, Appaso, Tirumala, Bhaskar, Nisha, Kumari
Transforming Battery Management: The Role of Artificial Intelligence and Machine Learning in Digital Twin Technologies2026-26-03821/16/2026
The traditional Battery Management System (BMS) faces certain limitations in fully utilizing battery capacity and performance during the long cycle life operation of Electric Vehicles (EVs). These constraints include limited real-time data collection, low processing speed, lack of predictive maintenance, and very limited accuracy in predicting health and degradation chemistry. A Battery Digital Twin (BDT) can effectively address these limitations of the BMS. Battery Digital Twins (BDT) can be viewed as a cyber-physical system comprising four key elements: virtual representation, bidirectional connection, simulation, and connection across the life cycle phases of an EV battery. The performance of a Li-ion battery largely depends on the cathode chemistry, component design, and operating conditions. The battery should be manufactured in a manner (such as cylindrical or prismatic cell) that prevents explosion, leakage, and gas generation inside the battery. To enhance the performance and
Chaturvedi, Vikash, M, Venkatesan, Lanke, Siddhi, Subramaniam, Anand, Karle, Manish, Pandit, Rugved, Gupta, Drishti, Karle, Ujjwala Shailesh
Automated Parameterization of CAE Model Based on Test Data Response Bands2026-26-04361/16/2026
This study presents a data-driven approach aimed at enhancing the correlation between physical test data and Computer-Aided Engineering (CAE) simulations, with an emphasis on adapting the standard CAE model's response to minimize any gaps relative to the response of a given test specimen. Leveraging historical test data, machine learning techniques are used to categorize responses into distinct bands, effectively capturing the inherent variability observed in real-world scenarios. This categorization step recognizes patterns across a wide range of test data, forming the foundation for closely matching and adapting CAE models to new, unseen hardware data. In typical automotive simulation workflows, tuning a standard CAE model to match new hardware test data involves iterative parameter adjustments and simulations. This process can be time-consuming and often lacks predictive insight into the necessary modifications. The approach developed in this study addresses this challenge by
Khopekar, Maria, Arya, Bibhu, Sridhar, Raam, Mohan, Pradeep, Kurkuri, Mahendra
Multi-Model Analysis of O-Ring Compression and Relaxation Force Predictions Using Machine Learning and CAE Simulations2026-26-04661/16/2026
O-rings play a critical role in ensuring leak-proof seals in a wide range of engineering systems. Accurate prediction of their compression and relaxation behavior under various material and geometric configurations is essential for optimal design and reliability. This study presents an analysis of machine learning techniques to predict two key performance outputs, compression force and relaxation force (after 10 minutes) trained on computer-aided engineering (CAE) simulation data. The experimental setup was represented in CAE simulation and the results were compared with experimental data conducted at ZF test facilities. Simulation results correlated well with the experimental data (deviation was less than the 5%). To create a dataset for training machine learning (ML) models, realistic ranges for the input parameters such as hardness and geometrical parameters were determined, and simulation data were generated using design of experiments (DOE). Multiple ML models were developed and
Kosgi, Durgaprasad, Alva, P Pancham, Dangeti, VenkataKrishna Pavan
Comparative Assessment of Lithium-Ion Battery Recycling Strategies: Unveiling Challenges, Optimizing Benefits, and Exploring Future Potential.2026-26-02011/16/2026
This research paper offers a comprehensive evaluation of lithium-ion battery recycling methods, tracing the entire journey from global demand to the practical challenges and solutions for sustainable battery recycling. It starts with the analysis of worldwide LIB demand growth alongside the exponential growth in volumes of spent batteries and recycling rates. The study focuses on the imbalance in production and recovery of critical battery components and its environmental and economic effects. The paper then systematically examines six major recycling methodologies: mechanical, pyrometallurgical, hydrometallurgical, biotechnological, direct, and ion-exchange recycling. It goes into detail about their advantages, limitations, and roles in maximizing the recovery of valuable metals such as lithium, cobalt, and nickel. Traditional techniques like hydrometallurgical and pyrometallurgical methods, and emerging approaches including bioleaching and ion-exchange, are evaluated for their
Jain, Gaurav, PREMLAL, P, Pathak, Rahul, Gore, Pandurang
Standardized design principles and emerging opportunities for AI-integrated CAD systems.2026-26-06321/16/2026
Artificial Intelligence (AI) in the automotive industry is growing and transforming in different segments of the industry. Still there is a significant gap persisting in the standardization of design principles and the incorporation of manufacturing constraints in system. However current development in AI CAD systems is isolated and non-parametric way, in contrary the conventional way of CAD methodology is knowledge based and systematic parametric steps which is agile to the iterative improvement. Hence it will be challenging in integration and adoption of these AI CAD systems in the well-established product development cycle. The research focuses on identifying the scope of AI integration which includes generative design, automated error detection, and design pattern-dependent learning systems, but also stresses the importance of standardized policies to address fundamental questions of system coherence, uniformity, and broad applicability. This research paper studies the adoption of
Shaikh, Taha Mukhtyarmiya, Harel, Samarth, Kumar, Akarsh, Venkitachalam, Muthukumar, SHAH, BHUMIKA, Chakraborty, Pinka
A Scalable ArTOP-Based In-House Toolchain for Automated AUTOSAR Configuration2026-26-06821/16/2026
The rising software complexity in Automotive industry demands reusable, hardware-agnostic development frameworks. AUTOSAR (Automotive Open System Architecture) provides a standardized, scalable ECU software architecture but cost-effective tooling and modern workflows are critical for broad adoption and competitiveness. One such area is for AUTOSAR configuration and authoring of Autosar architecture. Current solutions include commercial offerings which are often built on top of ARTOP (ArTOP is an eclipse-based ecosystem maintained by AUTOSAR consortium) and open-source python implementations. Commercial tools are prohibitive in cost, have complicated development workflows, are difficult to automate and lack quick integration with other tools. Python-based solutions are often community driven with small developer teams and face many challenges. These tools are not mature enough, have staggered development, security concerns, liability issues, lack of approvals etc. These libraries do not
Daware, Kartik, Garg, Mudit, Pasupuleti, Raju
Machine learning model to predict Chest Injury during Thorax Impact on Human Body Models2026-26-06651/16/2026
In recent years, virtual models have been handy in predicting potential injury risk to occupants in vehicle crashes. Virtual models offer detailing of occupant anthropometry and closest possible bio-fidelity over existing test devices. This study focuses on the assessment of chest deflections in frontal thorax impacts using virtual human body models of a few anthropometries and transforming the assessment of injuries of broader range of anthropometries (population). The study utilizes machine learning models to enable injury assessment across a broader range of body types. The study has taken a standard test scenario (Kroell load case) having frontal blunt thoracic impact. The load case is replicated in LS-DYNA and finite element human body models (HBMC) of 05th Female and 50th Male are used. The simulation setup replicates the boundary conditions as per literature, the responses necessary to measure chest deflection were validated & recorded. The outputs from the simulations and data
Sridhar, Raam, Arya, Bibhu, Divakar, Prajwal, R, Udhaya Kumar, Bhutki, Prasad, Kumar PhD, Devendra, Kurkuri, Mahendra, Mohan PhD, Pradeep
Video 2 Scenario: A Novel Method to Generate Edge Case Traffic Scenarios from Dashcam Videos2026-26-06661/16/2026
This paper presents Video 2 Scenario, a systematic and cost-effective method to generate edge case scenarios derived from road events captured in dash camera videos. Current methods for obtaining real-world scenarios rely on supplementary sensor data like LiDAR, GPS and radar, which require expensive sensor setups for data collection. Other scenario generation methods rely on manual effort by human experts, which can be laborious and time-consuming. Our proposed methodology utilizes state-of-the-art deep learning models to extract vehicle trajectories, road layouts and other scenario-specific elements and reconstructs a scenario in the format of simulation scripts. The scripts are then refined by replaying the scenario multiple times in a simulation with attached virtual sensors to minimize the discrepancy between the subsequently obtained simulated video data and the original dash camera video data. The refined simulation scripts ensure that critical edge case scenarios are replicated
Ramalingam, Surya Prasath, Priyadarshi, Marut, Khan, Parvej, Kumar, Vaibhav, Pandey, Gaurav, Lohani, Bharat
Self-supervised diagnostics & prognostics framework for predictive maintenance in electric vehicles2026-26-06731/16/2026
Predictive maintenance is essential for enhancing the reliability, safety, and operational efficiency of connected vehicles. However, traditional supervised learning approaches for fault prediction depend heavily on large, labeled datasets of failure events, which are often scarce and expensive to obtain in real-world automotive environments. This paper proposes an original self-supervised anomaly detection framework that predicts emerging failures using only healthy operational data, eliminating the need for explicit failure labels. The approach employs self-supervised pretext tasks, such as masked signal reconstruction and future telemetry prediction, to learn robust representations of normal multi-sensor behavior (e.g., temperature, pressure, current, vibration). An unsupervised anomaly detection model then identifies deviations from these learned patterns. This integrated with data-informed predictive models enable early detection of faults across critical subsystems such as
Kumar, Pankaj, Deole, Kaushik, Hivarkar, Umesh
Transforming Headlamp Levelling Compliance Testing Through AI-Powered Predictive Modelling2026-26-06601/16/2026
The process of validating headlamp levelling compliance as per regulatory standards involves physical testing with various vehicle loading conditions. This traditional method is labour-intensive, time-consuming, and consumes significant resources. There is a need for a predictive solution that can simulate and validate headlamp levelling tests virtually, thereby reducing dependency on physical trials. Headlamp levelling compliance is a critical regulatory requirement to ensure optimal visibility and safety under varying vehicle loading conditions. This paper presents an Artificial Intelligence and machine learning-based (AI / ML) solution to simulate headlamp levelling tests virtually/Digitally by using Historical Headlamp designs from past vehicles. By leveraging historical test data and developing regression models, the system predicts headlamp height and dipped beam height for both unladen and laden conditions. Key vehicle parameters such as tire load, overhang, deflection, and
Mandloi, Prince, Joshi, Vivek S., GHANWAT, HEMANT, Ugale, Anand, Munda, Rohit, GHAN, PRAVIN
Digital Twin for Motor & MCU incorporating Deep learning AI-ML algorithms for intelligent prognostics for Electric vehicles2026-26-04791/16/2026
As electric vehicles (EVs) become more advanced, so ensuring the reliability of critical components like the motor and Motor Control Unit (MCU) is essential. This paper presents a digital twin model designed to predict failures in motor and MCU components using machine learning. The approach focuses on detecting early signs of failure through real-world data and advanced analytics. We collected thermal and performance data from field vehicles, capturing both normal (healthy) and abnormal (faulty) operating conditions. Using this dataset, we developed and trained an Auto Encoder-based machine learning model that learns what “normal” looks like and flags deviations as potential issues. One key outcome of this study is the successful early prediction of Insulated Gate Bipolar Transistor (IGBT) degradation, where the system identified subtle behavioral changes long before any visible failure symptoms appeared. This digital twin acts as a virtual replica of the physical components
Joshi, Pawan, Pandey, Suchit, KONDHARE, Manish, UPADHYAY, ABHAY, JaganMoahanarao, Vana, Tank, Prabhu
Tractor Cabin Structure Borne Noise Reduction in the Virtual Environment2026-26-01011/16/2026
In recent days, cabin variants in the tractor are preferred by the farmers for the Coziness and longer field hour operation with less fatigue. Noise perceived by customer is the most important factor taken into account during the design stage, as it’s directly linked with operator’s comfort. Observed noise levels has to be within the defined limits as per national/international standards Overall cabin noise levels is contributed by the structure borne noise below 630 Hz. Structure borne noise is the noise typically radiated by the door, roof, windshield, floor, fender and structure assembly due to the engine excitation through the transmission housings and backstories. This paper depicts the process of tractor cabin structure borne noise prediction in the virtual environment. Firstly, Engine bearing loads and axle bearings has been extracted in the virtual stage from the vehicle level driveline model using commercially available MBD software. The finite element (FE) model of the cabin
gunasekaran, pandiyanayagam, K, Somasundaram, Chavan, Amit
A Deep Learning Approach for Dynamic Range Estimation in Electric Fleet Vehicles.2026-26-06751/16/2026
Keywords: Real-Time Range Prediction, Long Short-Term Memory (LSTM), Time Series Modelling, Deep Learning, Vehicular Telemetry, CAN Data, Range Anxiety Accurate and real-time range estimation is a critical enabler for efficient energy management and operational planning in Commercial Electric Vehicles (CVEVs). This project presents a robust deep learning-driven framework for vehicle range prediction, leveraging a Long Short-Term Memory (LSTM) network trained on high-frequency powertrain CAN data obtained from CVEVs operating across diverse driving environments. The dataset comprises key dynamic and environmental parameters—including State of Charge (SOC), vehicle speed, acceleration, and temperature—captured during real-world operations. Extensive data pre-processing techniques were applied to ensure signal integrity, including high-resolution timestamp synchronization, noise filtering, and imputation of missing values, thereby establishing a reliable foundation for time-series
surlekar, Samarth, Ghorpade, Swapnali, Kalghatgi, Bhakti
Data-Driven Prediction of Global Automotive Trends: Forecasting Fuel Economy and CO₂ Emissions Using Machine Learning2026-26-06431/16/2026
In the pursuit of cleaner transportation and environmental sustainability, the global automotive industry is undergoing rapid transformation. This study leverages historical multi-decade vehicle data to develop predictive models for fuel economy and CO₂ emissions across a wide range of vehicle technologies. By utilizing advanced machine learning algorithms in Python and integrating insights through Power BI visualizations, the project identifies key correlations between vehicle attributes—such as weight, powertrain, and footprint—and their environmental performance. Results highlight the increasing impact of electric vehicle adoption, hybridization, and light weighting on overall emissions reduction. These insights help forecast the direction of fuel economy standards, emission patterns, and technology shifts across manufacturers and vehicle types. Beyond technical predictions, the study offers a decision-support framework for global policymakers, automotive designers, and
Hazra, Sandip, Tangadpalliwar, Sonali, Hazra, Sanjana
Conceptual Framework for Real-Time Battery Health Estimation in Automotive Digital Twins Using Reinforcement Learning2026-26-06581/16/2026
The explosive growth of electric vehicles (EVs) calls forth the need for smart battery management systems that can perform health monitoring and predictive diagnostics in real-time. The conventional battery modeling methods mostly do not cover the complicated, dynamic behaviors coming from different usage patterns. The study outlines a structure that would use Reinforcement Learning (RL)-based AI agent as a part of the Battery Electrical Analogy (BEA) simulation platform. With the help of the AI agent, different health parameters such as State of Health (SOH), State of Charge (SOC), and the signs of early thermal runaway can be predicted in real-time. The suggested design takes advantage of the simulation-based approach to have the agent learn and utilizes a decentralized cloud architecture suitable for scaling and reducing the response time. The RL agent performs an essential role in the process by tagging along with the continuous learning and the adjustment of the battery conditions
Pardeshi, Rutuja Rahul, KONDHARE, MANISH, Sasi Kiran, Talabhaktula
Algorithm development for AI driven Path planning using Augmented reality for navigating in Indian complex road scenarios2026-26-00381/16/2026
The road infrastructure in India has complex navigational challenges, with 82% roads being unstructured according to Ministry of Roads Transports and Highways (MoRTH). Decision-making becomes a challenge for drivers in unpredictable environments such as narrow roads or sometimes flooded roads and heavy traffic. In this paper, an augmented reality (AR) driver assistance algorithm has been proposed that improves awareness to safely maneuver in these conditions.The methodology for development and validation of this AR algorithm contains multiple steps. Firstly, extensive data collection is conducted using real time capture and benchmark datasets like Berkeley Deep Drive (BDD) and Indian Driving Dataset (IDD). Secondly, A machine learning model consisting of the collected data are labelled and trained to rightly identify the complex scenario. In third step, an AR navigation algorithm is developed, integrating these models to accurately detect floods and estimate road widths and provide
Anandaraj, Prem Raj, Sivakumar, Vishnu, Thanikachalam, Ganesh, L, Radhakrishnan, Motoki, Yaginuma, Selvam, Dinesh Kumar
Master Complexity of Version Management in SDM by AI-Driven Assistance2026-26-04391/16/2026
Simulation-driven product development involves numerous CAE model iterations, where each version represents a critical difference. Usually, these multiple model versions are generated by hundreds of simulation engineers working in teams distributed across the globe, making functional collaboration a key for effective product development. To manage vast amounts of CAE data generated by engineers working simultaneously on a project, it is imperative to have a robust version management system to track changes in the CAE data. A robust version management is the backbone of an effective simulation data management (SDM) system. It involves capturing and documenting model changes at every design iteration. An accurate documentation of the model changes is crucial as it helps in understanding the model evolution and collaboration among engineers. However, documenting is usually considered a boring and tedious task by many engineers. This often leads to bad change documentation, which in turn
Thiele, Marko, Sharma, Harsh
Optimizing System Design performance and its accuracy using Feature Importance2026-26-04211/16/2026
With continuous improvements in model accuracy and computational performance, system simulation can be seamless integrated into development tools and methodologies as a "virtual test bed" in synergy with experimental tests. However, virtual vehicle and powertrain thermal models still face significant obstacles in realizing their full benefits. One common challenge is modeling accuracy and its anomalies, particularly in models provided by internal automotive companies and external suppliers. The models developed for vehicle thermal management and its systems have been created using specific software, making it difficult to verify the accuracy of the models across different systems and their performance. Integrating different systems in a Co-simulation environment has affected the overall accuracy of the Vehicle Simulation Platform model. This situation significantly reduces the return on investment in model development. To address these issues, we propose deploying feature importance
Srinivasan, Rangarajan, SARAPALLI RAMACHANDRAN, Raghuveeran, ASHOK BHARDE, Pooja, SARAVANAN, Vivek
Measurement of Acoustic Sensitivity of Automotive Drum Brake to Multiple Faults2026-26-05551/16/2026
Multiple faults can occur over time in drum brakes, which are commonly used on the rear of heavy vehicles and two-wheelers, compromising passenger safety and ergonomics. Given the recent success of artificial intelligence in fault diagnosis, it can be applied effectively to automotive brakes during operation when combined with advanced measurement techniques. However, monitoring the brake health during operation is challenging due to the rotational motion of components and space constraints for sensors. Consequently, this study investigates the use of measured radiated noise from drum brakes (instead of vibrations) to detect faults. A test bench mimicking typical on-road braking dynamics has been developed to operate in a controlled environment with pre-determined fault levels. This setup measures acoustic signals during transient brake events. Acoustic responses from several experiments that imitate three common brake fault conditions are measured. The acoustic sensitivity to each
Shripad, Kumar Milind Rewanand, Yella, Akash, Sundar, Sriram
Prediction of Trimmed BIW Noise transfer function based on BIW response using Machine Learning2026-26-06401/16/2026
This research focuses on the prediction of the Noise Transfer Function (NTF) in a trimmed Body-in-White (BIW) structure by leveraging machine learning technique considering BIW-level NTF, dynamic stiffness at engine mount attachment points, and Vibration Transfer Function (VTF). The study employs an iterative methodology, adjusting the thickness of various BIW panels to analyse their impact on noise transfer characteristics. Frequency response functions are generated at the BIW attachment points to assess the effectiveness of these modifications. By systematically increasing and decreasing panel thicknesses, and stiffness at the attachment point and trimmed BIW level responses the study aims to optimize the acoustic performance of the BIW structure. The findings provide critical insights into the relationship between structural modifications and noise transfer, offering valuable guidance for automotive design and engineering during early design stage to reduce the development time.
Kulkarni, Prasad Ramesh, Bijwe, Vilas, Kulkarni, Shirish, Sahu, Dilip, Inamdar, Pushpak
Application of DOE Techniques to Identify Factors Influencing Gearbox Synchronizer Wear and Life.2026-26-04891/16/2026
Gearbox synchronizers are essential components for ensuring smooth gear shifting in both manual and automated manual transmissions. The lifespan of a synchronizer is influenced by a complex interplay of several factors, including synchro geometry, lubrication, friction material properties, clutch design parameters, powertrain vibrations, and various operating conditions. These factors contribute to the wear and degradation of synchronizers over time, affecting the overall performance and durability of the transmission system. This paper investigates the influence of various design and operational parameters on the synchronizer's life using a Design of Experiments (DOE) approach. A comprehensive experimental plan was meticulously developed to systematically vary key factors such as synchronizer friction material, the number of cones, lubricant viscosity, and clutch stiffness, among others. The resulting experimental data was analyzed through statistical methods, particularly regression
Jagtap, Amol Shivaji, Koulage, Dasharath, Rudramath, Sagar
Early Prediction of CNG Filling Time using Artificial Intelligence for Design Optimization2026-26-06621/16/2026
The CNG fuel system is a critical component in vehicle development and typically includes large-volume gas cylinders (up to 800 liters). As these vehicles often operate over long distances daily, frequent refueling is necessary, making a short gas filling time highly desirable. Currently, CNG filling time is evaluated after prototype development to determine the refueling duration. If the filling time is excessive, design changes to the fuel system are required, leading to significant cost and time penalties and causing delays in the vehicle product development cycle. By integrating AI/ML technology, a mathematical model has been developed based on various influencing parameters. This model will predict CNG gas filling time across all commercial vehicle platforms at the initial Zero design release gateway. This early prediction will enable further optimization to improve gas filling time. The goal of this research work to optimising the filling time for various platform before physical
Choudhary, Aditya Kant, Petale, Mahendra, Dutta, Surabhi, Bagul, Mithilesh
Advanced Indirect Tire Health Monitoring System Using Transfer Learning for Predictive Maintenance2026-26-06701/16/2026
Tire wear is a critical factor in vehicle safety and operational reliability, with worn-out tires contributing to nearly 30% of highway blowouts and a significant portion of tire-related accidents. Traditional tire wear indicators—such as tread wear bars or mileage-based replacement schedules—lack precision, do not offer real-time feedback, and often fail to prevent hazardous situations caused by uneven or rapid tire degradation. These shortcomings underscore the urgent need for intelligent, data-driven solutions in predictive tire maintenance. While earlier efforts have focused on physical models or basic machine learning algorithms, their limited adaptability and poor generalization across platforms reduce effectiveness in real-world deployments. This paper presents a Machine Learning-based Tire Wear Prediction System, designed for both cloud-based analytics and on-vehicle edge deployment. The system predicts individual tire wear using CAN-based inputs, supported by a transfer
Imteyaz, Shahma, Iqbal, Shoaib
Chest Deflection mapping between HBM Connect® 50M and Humanetics ATD models for frontal impact2026-26-00101/16/2026
Objective Human Body Models (HBMs) are increasingly recognized by consumer protection agencies as essential tools for evaluating vehicle safety, complementing the use of traditional Anthropomorphic Test Devices (ATDs). However, HBMs are new for product development and pose challenges in connecting them with the ATDs. These challenges can be overcome if a link is established for the injury metrics between HBMs and ATDs. The study aims to develop a chest deflection mapping function between the HBM Connect® 50M (HBMC-50M) and two ATD models: THOR-50M and Hybrid III 50M for thoracic assessment in frontal impact. Method Several frontal chest loading scenarios were selected from the HBM4VT qualification catalogue (Euro NCAP technical bulletin CP 550), including hub impacts and table-top tests. Matched-pair LS-DYNA simulations were conducted with HBMC-50M and the ATD models recording peak chest deflections for developing the mapping function. In the HBMC-50M model, deflection outputs were
Velmurugan, Gopinath, Kumar PhD, Devendra, R, Udhaya Kumar, Kulavi, Pradeep, Soni, Anurag, Tejero de la Piedra, Ricardo, Fu, Stephen, Long, Teng, Arora, Tushar, Jagadish, Renuka, Shah, Chirag
A Mathematical Modelling Approach Based on Artificial Neural Network for Estimation of Key Parameters in Internal Combustion Engine2026-26-06591/16/2026
In a conventional powertrain driven by Internal combustion (IC) engines, there are many key sensors such as intake manifold temperature sensor, exhaust temperature sensor, Nitrogen oxide (NOx) sensors etc. to monitor engine performance and emissions. For enabling various diagnostics strategies and engine monitoring, virtual sensors or modelled values of such sensors play an important role. Conventional strategies incorporate the use of regression models, map-based models and physics-based models. There are a few drawbacks with conventional models in terms of accuracy and model calibrations efforts. Data driven models or neural networks have fairly better accuracy and reliability for estimating complex parameters. Representing the neural network with a mathematics-based model would help to eliminate drawbacks associated with conventional modelling approach. The proposed methodology uses artificial intelligence technique called artificial neural network (ANN) for estimation, for example
Jagtap, Virendra Shashikant, Shejwal, Sanket, Mitra, Partha
Towards AI-Driven Automated Driving Systems: Homologation perspective2026-26-01331/16/2026
This paper examines the challenges and opportunities in homologating AI-driven Automated Driving Systems (ADS). As AI introduces dynamic learning and adaptability to vehicles, traditional static homologation frameworks are becoming inadequate. The study analyzes existing methodologies, such as the New Assessment/Test Methodology (NATM), and how various institutions address AI incorporation into ADS certification. Key challenges identified include managing continuous learning, addressing the "black-box" nature of AI models, and ensuring robust data management. The paper proposes a harmonized roadmap for AI in ADS homologation, integrating safety standards like ISO/TR 4804 and ISO 21448 with AI-specific considerations. It emphasizes the need for explainability, robustness, transparency, and enhanced data management in certification processes. The study concludes that a unified, global approach to AI homologation is crucial, balancing innovation with safety while addressing ethical
Lujan Tutusaus, Carlos, Hidalgo, Justin
AI-Driven Stress and Fatigue Life Prediction of Engine Connecting Rods using Geometric Deep Learning2026-26-06331/16/2026
The connecting rod is one of the critical components of the engine which is subjected to complex cyclic loading, demanding precise evaluation of stress distribution and fatigue life. Conventional Finite Element Method (FEM)-based simulations, while accurate, are computationally expensive and time-intensive—especially when iterating through multiple design variations. To address this, we present an AI/ML-based predictive framework using Graph Neural Networks (GNNs) and Geometric Deep Learning (GDL) techniques to efficiently predict stress and fatigue life of the connecting rod. The methodology involves representing FEM mesh data as graph structures, where nodes and edges capture geometric and physical relationships. GNN architectures, including MeshCNN and PointNet++, are trained on historical simulation data to learn spatial dependencies and nonlinear mappings between geometry, load conditions, and resulting stress/life outputs. The framework significantly reduces computation time
Pathan, Mohammed Shakil, K, Karthikeyan, pilla, sashanka, S Kangde, Suhas
Predictive Modelling of Battery Behavior for Enhanced Energy Management in Electric Vehicles2026-26-03851/16/2026
As electric vehicles (EVs) and smart energy systems continue to evolve, optimizing energy usage and ensuring system reliability has become critical. However, battery pack modelling remains challenging due to its highly dynamic and nonlinear behavior due to the temperature variations, aging, and variable load conditions. This paper presents a 1D digital twin framework designed for predictive monitoring of the battery parameters, including State of Charge (SoC), terminal voltage, temperature, and an indicative assessment of State of Health (SoH). The battery is modelled using a Thevenin equivalent circuit that covers both steady-state and dynamic electrical characteristics. Key parameters such as open-circuit voltage (OCV), internal resistance, and dynamic characteristics were derived through pulse testing conducted in our laboratory. A thermal modelling is done to capture the heat generation and cell temperature. SoC estimation is carried out using the Coulomb counting method and
G, Ayana, Gumma, Muralidhar
Identification of productive and non-productive activities in agricultural machinery using ECU and GPS parameters through the application of machine learning techniques2026-26-01021/16/2026
Any agricultural operation (such as cultivation, rotavation, ploughing, and harrowing) includes both effective work and non-productive activities (like transportation, stops, and idling) in the field. Non-productive work can distort the actual load profile, fuel consumption, and emissions. In this project, a machine learning-based methodology has been developed to differentiate between effective operations and non-productive activities, utilizing data collected in the field from data loggers installed on the machinery. Measurements were conducted on various machines across the country in all major applications to minimize the influence of any individual sample's deviation and to account for variability in customer operating practices. Nine critical parameters (Engine Speed, Exhaust Gas Temperature, THECU2, VPEXGAS, FULRAT, Actual Engine Percentage Torque, LBEGR, POIL & GPS Speed) were selected after screening & analyzing more than 100 CAN and GPS parameters. The critical parameters
Maharana, Devi prasad, Gangsar, Purushottam, gokhale, Varun, Pandey, Anand Kumar
AI-Driven Digital Twin for Enhanced Suspension Assembly Testing in Automotive Vehicles2026-26-04451/16/2026
This paper presents the development and application of a digital twin for the suspension assembly of Automotive vehicle, a critical component in evaluating the performance and durability of automotive systems. The digital twin, a virtual replica of the physical suspension assembly, is created using advanced simulation techniques and physical test data integration. By leveraging machine learning algorithms, the digital twin provides a comprehensive and dynamic model that accurately reflects the behavior and performance of the physical system under various conditions. The primary objective of this study is to enhance the accuracy and efficiency of suspension assembly testing by enabling predictive maintenance, real-time monitoring, and optimization of test parameters. The digital twin allows for the identification of potential issues before they occur, reducing downtime and maintenance costs. Additionally, it facilitates the exploration of different test scenarios and conditions without
Sonavane, Pravinkumar, Patil, Amol
Machine Learning based Engine Mount NVH Life Health Monitoring System2026-26-03591/16/2026
Refinement for NVH (Noise, Vibration, and Harshness) performance, and long-term vehicle reliability are rapidly evolving in today’s automotive industry. Engine mounts play a central role in isolating powertrain-induced vibrations; their deterioration can significantly affect cabin comfort, powertrain integrity, and customer satisfaction. Prior work in this area has primarily focused on direct mount sensors and physical inspection at service centre after failure. While effective in controlled environments, such methods are not scalable, add system complexity, and increase vehicle cost due to sudden breakdowns—challenges that restrict practical OEM deployment. This paper introduces a novel indirect health monitoring method that leverages a driver seat rail-mounted accelerometer to capture driver specific vibrational responses. By analyzing these signals using machine learning models and domain-specific analytical features, engine mount health is inferred without requiring sensors on the
Iqbal, Shoaib, Dusane, Mangesh
A Novel Wide Range Machine Learning Polynomial Regression Formulation for Enhanced Load Cell Calibration Accuracy2026-26-05671/16/2026
Calibration of measuring instruments is of utmost importance in the field of metrology. It is a mandatory pre-requisite for establishing the fidelity of the measurements as well as to lend confidence. Even more critical is the requirement for the master equipment deployed to calibrate the devices in use. This entails that high accuracy needs to be guaranteed in the calibration process, and that the uncertainty be quantified precisely. The widely used conventional least squares polynomial regression formulation for load cell calibration is based on the non-normalized residual, which is the difference between the measured and master values. The nature of this formulation is such that it imparts more weightage on measured values at higher ranges resulting in good accuracy. However, there is a limitation of this same formulation that results in lesser accurate fit at lower values especially if the instrument is to be used in operation over a wide range including lower ranges of the
S Thipse, Yogesh
Advanced Defect Detection Techniques Using AI, CAD Data, and Computer Vision2026-26-06441/16/2026
In area of modern manufacturing, ensuring product quality and minimizing defects are utmost important for maintaining competitive advantage and customer satisfaction. This paper presents an innovative approach to detect defect by leveraging Artificial Intelligence (AI) models trained using Computer-Aided Design (CAD) data. Traditional defect detection methods often rely on physical inspection, which can be time-consuming and prone to human error. The conventional method of developing an AI model requires a physical part data , By utilizing CAD data, the time to develop an AI model and implementing it to production line station can be saved drastically . This approach involves the use of AI algorithms trained on CAD models to detect and classify defects in real-time. The field trial results demonstrate the effectiveness of this approach in various industrial applications, highlighting its potential to revolutionize defect detection in manufacturing.
Kulkarni, Prasad Ramesh, Sahu, Dilip, Joshi, Chandrashekhar, Khatavkar, Akshay, Poddar, Shivani, Deep, Amar
Computer Vision Aided Methodology for Steering Position Detection2026-26-06311/16/2026
Steering angle is a critical parameter in vehicle dynamics, traditionally measured using Steering Angle Sensors (SAS). However, external sensors used for development testing often pose challenges due to their bulk, installation complexity, and additional costs. With the upcoming regulatory requirements related to driver drowsiness detection, vehicles will incorporate cameras in near future. This study explores an innovative approach to repurpose the same camera for steering angle measurement, eliminating the need for dedicated sensors and reducing costs significantly. The proposed method utilizes colored tapes or visible markers affixed to the steering wheel, with a strategically placed camera capturing steering movement. Depending on requirements, various types of cameras (including thermal) can be deployed at different positions (front, back, or side). The vehicle's ECU employs an advanced object detection algorithm to track the movement of these markers, subsequently calculating the
TP, Amal, Priyadasini, Shishira
Items per page:
1 – 50 of 1396