Robert Bosch (India)

<div class="section abstract"><div class="htmlview paragraph">Software-defined vehicles are those whose functionalities and features are primarily governed by software, thus allowing continuous updates, upgrades, and the introduction of new capabilities throughout their lifecycle. This shift from hardware-centric to software-driven architectures is a major transformation that reshapes not only product development and operational strategies but also business models in the automotive industry. An SDV operating system provides the base platform to manage vehicle software and enable those advanced functionalities. Unlike traditional embedded or general-purpose operating systems, it is designed to meet the particular demands of modern automotive architectures. Reliability, safety, and security become crucial because even minor faults may have serious consequences. Key challenges to be handled by the SDV OS include how to handle software bugs, perform real-time processing, address functional safety and SOTIF compliance, adhere to regulations, minimize attack surface exposure, and protect against remote access and data breaches. This is achieved via sound architectural principles, including a CSM for fine-grained access control, a lean and minimal kernel to reduce vulnerabilities, secure and efficient inter-process communication, and user-level drivers to provide better fault isolation. The key novelty of this approach rests on the fact that it uses open-source kernels, libraries, and tools that guarantee flexibility, clarity, and community-driven innovation. It provides a flexible runtime environment and OS-level isolation using virtualization, safe hardware sharing, and adherence to safety standards to set up the SDV OS as a resounding, secure, and future-ready base for next-generation automotive systems.</div></div>

<div class="section abstract"><div class="htmlview paragraph">As the automotive industry transitions toward software-defined vehicles and highly connected ecosystems, cybersecurity is becoming a foundational design requirement. A challenge arises with the advent of quantum computing, which threatens the security of widely deployed cryptographic standards such as RSA and ECC. This paper addresses the need for quantum-resilient security architectures in the automotive domain by introducing a combined approach that leverages Post-Quantum Cryptography (PQC) and crypto-agility. Unlike conventional static cryptographic systems, our approach enables seamless integration and substitution of cryptographic algorithms as standards evolve.</div><div class="htmlview paragraph">Central to this work is the role of Hardware Security Modules (HSMs), which provide secure, tamper-resistant environments for cryptographic operations within vehicles. We present how HSMs can evolve into crypto-agile, quantum-safe platforms capable of supporting both hybrid (RSA/ECC + PQC) and fully post-quantum deployments—ensuring secure transitions without requiring hardware replacement. The novelty of this work lies in the design and validation of a first-of-its-kind operational prototype that supports current cryptographic standards (RSA/ECC) and is engineered for plug-and-play migration to PQC.</div><div class="htmlview paragraph">Our architecture ensures long-term security while minimizing operational disruption and costs. Using a systematic architectural methodology, we integrated both software- and hardware-based HSMs and evaluated their performance under hybrid cryptographic conditions. Key performance metrics such as latency, key negotiation time, and re-keying efficiency demonstrate that crypto agility can be achieved with minimal overhead, confirming its feasibility for real-world deployment.</div><div class="htmlview paragraph">There is an urgent need to adopt quantum-safe and agile security practices today, as vehicles manufactured now will remain in service long after quantum computers become practical. By embracing crypto-agile designs, automakers can mitigate long-term risks and ensure resilience against future cryptographic threats. This paper provides both a technical roadmap and a working prototype demonstration to guide the automotive industry toward a secure, quantum-resilient future.</div></div>

Vitreous hemorrhage is a serious ocular condition that requires accurate and timely diagnosis. Traditional methods for identifying affected areas in fundus images have limitations in precision and reliability. There is a need for advanced techniques to enhance the detection of vitreous hemorrhage and improve patient outcomes. Detecting hemorrhages in color fundus images is difficult due to elements such as noise, feature intensity, and overlapping areas. In particular, the variation in feature intensity between different sections of the retina and lens leads to uniform (homogeneous) characteristics. This study introduces a Feature Intensity Vision Transformer (FIVT) model utilizes vision transformer technology to detect high-intensity regions within segmented patches based on minimal pixel distributions. The model follows a systematic pipeline involving input image pre-processing, patch extraction (1 × 1-32 × 32), feature intensity estimation using the gray-level co-occurrence matrix (GLCM), and Vision Transformer-based intensity classification to produce a precise region detection mask. Feature distribution is analyzed to identify maximum intensity areas. High and low-intensity regions are classified, and pixels with significant feature content are matched with trained inputs. The model is trained using external images and high-feature regions until accuracy stabilizes. We utilize the "exudate-hemorrhage-health" retinal image dataset to evaluate the proposed FIVT. This dataset comprises over 3,000 fundus color images, which include both internal and external references. Out of these, 536 images showing infections are employed to examine the infected areas. For the training process, 3,000 images are categorized into healthy, hemorrhage, and exudate groups. Compared to other methods, the FIVT model achieved a 12.13% improvement in detection accuracy, a 12.55% increase in specificity, and a 12.32% enhancement in sensitivity, demonstrating robust performance across varying illumination and feature intensity conditions. The model effectively handles varying illumination and intensity distributions, improving the precision of vitreous hemorrhage localization. However, its performance can be further enhanced with larger, multi-institutional datasets and integration of 3D volumetric data. Future work will focus on clinical validation and adaptation of the framework for related retinal disorders such as diabetic retinopathy and macular edema.

<div class="section abstract"><div class="htmlview paragraph">The Exhaust Emission Control is a vital part of automotive development aimed at ensuring effective control of pollutants such as NOx, CO, and HC. The traditional method of calibrating emission control strategies is a highly time-consuming process, which requires extensive vehicle testing under a variety of operating conditions. The frequent updates in emission legislation requires a high-efficiency process to achieve a faster time-to-market. The use of Machine Learning (ML) in the domain of emission calibration is the need of the hour to proactively improve the process efficiency and achieve a faster time-to-market. This paper attempts to explores emerging trend of Machine Learning (ML) based data analysis that have improved the overall process efficiency of emission control calibration. The data generated by automated programs could be used directly in data analysis with minimal or no need for data cleaning. The Machine Learning (ML) models could be trained by historical data from relevant engine platforms to predict the output. The integration of Machine Learning (ML) models with automated measurement processes further enhances the process by enabling model-based calibration development. The use of automated programs and machine learning (ML) models could ensure high accuracy of the emission calibration data. This methodology could significantly reduce the need for volumes of measurements required for data analysis and calibration. This could further help in optimized usage of testing facilities, ultimately saving time and resources. A 70% overall savings in time and resources could be expected with the use of automation and machine learning models. This methodology also supports faster calibration development cycles that would be required for adhering to frequent legislative changes and achieving faster time-to-market.</div></div>

The long-term durability of the cathode catalyst layer continues to be one of the main obstacles for the widespread commercialization of proton exchange membrane fuel cells (PEMFCs) for transportation applications. While the investigation of PEMFC catalyst degradation typically focuses on Pt coarsening due to load/unload voltage cycling, this talk aims to discuss less thoroughly understood corner cases that are determined by system limitations. Specifically, H2/H2 short-stops and air/air starts will be discussed in the context of subscale tests under controlled conditions. The topics are discussed in 3 parts:Part 1 discusses trade-offs between degradation and recovery during H2/H2 short-stops with an emphasis on how the mechanistic understanding and quantification of stressors enables improvements in stack performance over lifetime while mitigating the negative effects of degradation. For this, the ECSA and H2/air performance decay is quantified for various air-starvation conditions, varying the upper potential limit, bleed down rate, and lower potential limit, and the effect of Pt-reduction on mass activity and high current density performance is evaluated.Part 2 discusses air/air start testing protocols typically used in literature, particularly the correlation between voltage losses and ECSA, Pt particle size growth, catalyst layer thinning, and homogeneity of aging. It is shown that the air/air start protocol resulted in a combination of comparatively mild carbon corrosion and severe Pt dissolution/redeposition that caused significant Pt size growth and Pt band formation. Based on these findings, an industry relevant alternative voltage-controlled protocol that combines the two main degradation mechanisms of air/air starts is proposed: triangle sweep cycles at 500 mV/s between 0.7 V and 1.4 V (H2/N2, 45ׄ°C, 100% relative humidity), thus enabling a more relevant catalyst/CCM development.Part 3 discusses air/air starts with preceding H2-protected shut-down, thus combining parts 1 and 2 with a focus on how an understanding of degradation allows Bosch to adapt both the materials selection and the operating strategy to maximize PEMFC performance and lifetime.

Abstract The Sulfate Removal Process (SRP) plays a vital role in offshore and onshore oil and gas operations by reducing sulfate concentrations in seawater before it is injected into reservoirs. This process helps prevent scale formation and mitigates reservoir souring caused by sulfate-reducing bacteria. The SRP typically includes a multi-stage filtration system—comprising coarse filtration, ultrafiltration (UF), and nanofiltration (NF). Among these, the UF stage, which targets the removal of fine particulates, has been repeatedly identified as a source of mechanical failure. Specifically plate distortion, bolt rupture, and membrane damage, all occurring under standard operating conditions which lead to reduce production capacity below 70%. Leveraging the innovative concept of Reality Based Prescriptive Predictive Maintenance (RB-PsM/PdM); a concept innovated by Arthur 1, a combined simulation and AI-driven approach was adopted to diagnose root causes and enable predictive maintenance. Computational Fluid Dynamics (CFD) was employed to analyze internal flow patterns, revealing pressure hotspots and non-uniform load distributions on the membrane. Simultaneously, Finite Element Analysis (FEA) evaluated the structural response of housing components, bolts, and support plates under both filtration and backwash loading. Fatigue life analysis further quantified the long-term impact of cyclic pressurization. Crucially, insights from these simulations were integrated with operational data using machine learning algorithms to develop a predictive maintenance model. This AI-enhanced model enables near-real-time monitoring and early failure prediction, significantly improving the reliability of the UF system. The study also proposes design upgrades —including an advanced sandwich plate made of Incoloy and Titanium, improved bolt grade and material, and rib-reinforced chamber sections—which collectively reduce cyclic stress concentrations and extend component life. As a result of the design enhancements and predictive maintenance model, the system's operational reliability significantly improved, leading to a production capacity increase of 12-18%. This translated into extended uptime, reduced unplanned shutdowns, and enhanced overall process efficiency. The novelty of this work lies in its integrated approach that combines high-fidelity physics-based simulations with artificial intelligence to both diagnose and proactively manage mechanical failures in critical filtration systems. Unlike conventional failure analysis that is reactive in nature, this study leverages the predictive power of AI trained on insights derived from Computational Fluid Dynamics (CFD), Finite Element Analysis (FEA), and historical operational data. By translating simulation outputs into features for machine learning models, a robust predictive maintenance framework was developed, capable of forecasting component failures—such as bolt loosening, bolt shear, gasket degradation, and membrane rupture—before they occur. This fusion of physics and data-driven intelligence not only enhances the diagnostic depth but also transforms maintenance strategies from time-based to condition-based, paving the way for smarter, more resilient sulfate removal systems in the oil and gas industry.

Decision making for the treatment of kidney failure in older adults should incorporate patient preferences. This discrete choice experiment (DCE) identified the health outcome preferences of older patients with chronic kidney disease (CKD) who are at elevated risk of kidney failure. DCE questionnaire development and application. Persons aged 65 years and older with CKD stage 4 or 5 treated in an outpatient clinic. Five questionnaire attributes: fatigue, life expectancy, level of independence, hospital admissions, and hospital visits. Patients' health outcome priorities (relative importance of attributes) and the trade-offs they were willing to make (marginal rates of substitution). DCE design based on qualitative research, literature review, and pilot testing (n=14). Data were analyzed using a logistic regression analysis, and interactions between age subgroups were explored using logistic regression models. The questionnaire was completed by 85 patients (mean age 77 years; 65% male). Fatigue had the greatest influence on patients' choices (relative importance 26%), followed by life expectancy (23%), hospital admissions (20%), level of independence (16%), and hospital visits (15%). To avoid fatigue, patients were willing to accept a 46% decrease in 3-year survival probability, 5 extra hospital admissions per year, or 86 extra hospital visits per year. Among patients aged≥85 years, avoiding hospital admissions was the top health outcome priority; among patients aged 65-74 years, it was the lowest-ranked attribute. Participants were required to return the questionnaire independently. Subgroups may have been too small to detect all significant differences. Fatigue and life expectancy were the top health outcome priorities of older people with advanced CKD. Avoiding hospital admissions was the top priority among the oldest patients. These findings inform the choice of patient-centered outcomes in research and clinical practice for older patients with CKD. This study explored what matters most to older patients with kidney disease when making decisions about treatments. We asked patients to choose between treatment options that varied in terms of fatigue, life expectancy, independence, hospital admissions, and hospital visits. The findings revealed that fatigue and life expectancy were the most important factors for patients aged 65 years or older. The subgroup of patients aged 85 years or older cared most about avoiding hospital admissions. These findings help us understand what older patients with kidney disease may care about most when deciding on treatment options, which may inform their treatment decisions.

Visual prompt-based methods have seen growing interest in incremental learning (IL) for image classification. These approaches learn additional embedding vectors while keeping the model frozen, making them efficient to train. However, no prior work has applied such methods to incremental object detection (IOD), leaving their generalizability unclear. In this paper, we analyze three different prompt-based methods under a complex domain-incremental learning setting. We additionally provide a wide range of reference baselines for comparison. Empirically, we show that the prompt-based approaches we tested underperform in this setting. However, a strong yet practical method—combining visual prompts with replaying a small portion of previous data—achieves the best results. Together with additional experiments on prompt length and initialization, our findings offer valuable insights for advancing prompt-based IL in IOD.

The advances in multi-modal foundation models (FMs) (e.g., CLIP and LLaVA) have facilitated the auto-labeling of large-scale datasets, enhancing model performance in challenging downstream tasks such as open-vocabulary object detection and segmentation. However, the quality of FM-generated labels is less studied as existing approaches focus more on data quantity over quality. This is because validating large volumes of data without ground truth presents a considerable challenge in practice. Existing methods typically rely on limited metrics to identify problematic data, lacking a comprehensive perspective, or apply human validation to only a small data fraction, failing to address the full spectrum of potential issues. To overcome these challenges, we introduce VISTA, a visual analytics framework that improves data quality to enhance the performance of multi-modal models. Targeting the complex and demanding domain of open-vocabulary image segmentation, VISTA integrates multi-phased data validation strategies with human expertise, enabling humans to identify, understand, and correct hidden issues within FM-generated labels. Through detailed use cases on two benchmark datasets and expert reviews, we demonstrate VISTA's effectiveness from both quantitative and qualitative perspectives.