The rapid progression of automotive technology through In-Vehicle Networks (IVNs) such as Controller Area Networks (CAN), Local Interconnect Network (LIN), FlexRay, Media Oriented Systems Transport (MOST), and Automotive Ethernet has significantly reshaped internal vehicle communication. This expansion in connectivity has led to new cybersecurity risks that directly affect passenger safety and the reliability of transportation infrastructure. These challenges highlight the need for resilient infrastructure for sustainable industrialisation. Many existing reviews in the IVN cybersecurity domain focus individually on specific protocols or attack categories and corresponding defense strategies rather than considering the entire framework. Thus, we conduct a Systematic Literature Review (SLR) on existing literature from 2010 to 2025 and examine cybersecurity developments related to IVNs. A structured taxonomy that integrates IVN protocols, major attack techniques, and datasets, along with cryptographic, conventional and Artificial Intelligence (AI)-based cybersecurity mechanisms, is presented. Our study adheres to PRISMA guidelines and provides a comprehensive and systematic analysis of the IVN cybersecurity domain. This provides a consolidated data synthesis of existing literature, offering deeper insights into current trends, research gaps, and emerging research directions in IVN cybersecurity. The SLR also emphasizes the transition from static, rule-based methods to adaptive and hardware-based solutions and identifies that the key trend is the shift toward adaptive cybersecurity mechanisms in heterogeneous IVN environments. Finally, the limitations in current solutions are also identified as research gaps, and the future research directions are outlined to strengthen the cybersecurity resilience of autonomous vehicles.
To enable efficient planning and operation of optical networks, it is essential to evaluate the Quality of Transmission (QoT) of lightpaths before they are established. The nonlinear effects in optical fibers cause deployed lightpaths to impact each other’s QoT. Therefore, reliable QoT estimation is essential, as it enables the assessment of lightpath feasibility prior to deployment and supports efficient network planning and optimization. Existing deep learning-based QoT estimation methods typically predict the Q-factor directly, which can make learning challenging under slowly varying network conditions. In this paper, we propose a Residual Transformer for Q-factor estimation that incorporates a novel temporal-mean residual prediction formulation. Instead of directly regressing the absolute Q-factor, the proposed model predicts the deviation from the temporal-mean Q-factor within the input window, and reconstructs the final prediction by adding the estimated residual to the baseline. By exploiting the strong temporal persistence of QoT measurements, the proposed formulation simplifies the regression task, enabling the model to focus on learning residual temporal dynamics rather than the complete Q-factor trajectory. The proposed framework combines this residual prediction formulation with a single-head Transformer encoder and is evaluated on a real-world dataset collected from Microsoft’s North American optical backbone network. Experimental results demonstrate that the proposed method achieves an MSE of 0.00296 and an MAE of 0.0318, outperforming nine standard baseline models, including a multilayer Long Short-Term Memory (LSTM) network, a multilayer perceptron (MLP), Support Vector Regression (SVR), a simple neural network (NN), a Gated Recurrent Unit (GRU), Temporal Convolutional Network (TCN), PatchTST, iTransformer, and TimesNet. Simulation results reveal that the proposed model consistently achieves outstanding performance across multiple channels in the dataset, with an MSE of 0.00296 and MAE of 0.0318, highlighting its suitability for advanced flexible optical networks. Additional experiments show that incorporating the proposed temporal-mean residual prediction formulation into GRU and LSTM models also substantially improves their prediction accuracy, indicating that the formulation is not limited to the Transformer backbone.
In the present study, the weldments were fabricated via friction stir welding (FSW). The dissimilar matrix metals of AA 7075 (T651) and AA 6061 (T6) with nanohexagonal boron nitride (h-BN) as reinforcement were the working materials considered. The joints were distinguished in terms of varied volume fraction (3, 5 and 7
In the last few years, the fusion of multi-modal data has been widely studied for various applications such as robotics, gesture recognition, and autonomous navigation. Indeed, high-quality visual sensors are expensive, and consumer-grade sensors produce low-resolution images. Researchers have developed methods to combine RGB colour images with non-visual data, such as thermal, to overcome this limitation to improve resolution. Fusing multiple modalities to produce visually appealing, high-resolution images often requires dense models with millions of parameters and a heavy computational load, which is commonly attributed to the intricate architecture of the model. We propose LapGSR, a multimodal, lightweight, generative model incorporating Laplacian image pyramids for guided thermal super-resolution. This approach uses a Laplacian Pyramid on RGB colour images to extract vital edge information, which is then used to bypass heavy feature map computation in the higher layers of the model in tandem with a combined pixel and adversarial loss. LapGSR preserves the spatial and structural details of the image while also being efficient and compact. This results in a model with significantly fewer parameters than other SOTA models while demonstrating excellent results on two cross-domain datasets viz. ULB17-VT and VGTSR datasets.
Abstract Single Point Incremental Forming (SPIF) offers high flexibility for producing complex sheet metal components. However, simultaneous high formability and high surface integrity remain problematic with aluminium alloys. This study investigates the SPIF process applied to AA-2024-O sheets where the goal is to maximize forming depth (Zmax) and at the same time reduce the roughness of the surface (Ra). Taguchi based mixed orthogonal array was used to test eight critical process parameters. Then eempirical models were developed using the experimental results. A constrained multi-objective optimization framework that included the Non-Dominated Sorting Genetic Algorithm II (NSGA-II) was used to find Pareto-optimal solutions. These were ranked using the MOORA decision-making method to identify a balanced and practically achievable parameter space. The optimal parameter combination resulted in surface roughness of 0.026079 and a forming depth of 48.11 mm. Close correlation with the predicted values was shown by verification experiments with less than 5 percent deviation. The proposed combined solution provides a reliable and experimentally validated framework on which the SPIF parameters to be used on AA-2024-O can be chosen, to enhance formability and surface quality.
