Abstract As urban populations continue to grow, managing and optimizing urban mobility has become increasingly complex, especially in multimodal transport interchanges. Accurate passenger flow measurement has therefore become essential for operators to mitigate congestion and improve service efficiency. This work proposes a scalable and flexible end-to-end system designed to accurately measure and track passenger flow in real-time. The system integrates a distributed network of Edge-AI sensor nodes with deep learning algorithms for local passenger detection and tracking, while a central processing server aggregates node outputs to derive flow counts. This approach overcomes the limitations of traditional single-sensor solutions by effectively handling occlusion and complex spatial configurations across multiple access points. Validated in a high-transited transport hub, results show that the system achieves accuracy rates between 94.03% and 99.30% even under crowded conditions with flow rates of 100 persons per minute, demonstrating its robustness and practical applicability in dynamic, high-density environments.

One of the most significant challenges faced by computer vision algorithm developers is effectively handling images captured in variable illumination scenarios, which may include poorly lit environments, object occlusions, reflections, or even partial visibility. The introduction of deep learning techniques has marked an unprecedented advance in segmentation, identification, and object reconstruction in such environments. This research presents a deep learning architecture based on a Conditional Generative Adversarial Network (cGAN) specifically designed for various grape berry segmentation varieties under any illumination conditions. It exploits a dedicated data collection for training and testing the cGAN segmentation model, that consists of 25,350 images of five grape berry varieties, divided into two sets: 20,350 for training and 5,000 for testing. To evaluate the impact of convolution kernel sizes on the segmentation process within the GAN architecture, six models were trained using kernel sizes of 2x2, 4x4, 6x6, 8x8, 10x10, and 12x12. Five metrics (accuracy, precision, recall, F 1 -score, and IoU) were employed to facilitate comparison of result with similar segmentation methods. The metrics have been calculated on a dataset with images in variable illumination conditions and in five specific scenarios with images categorized as: (a) good illumination, (b) poor illumination; (c) lack of illumination; (d) noisy image and (e) noisy and poor illumination. Among all the assessed model configurations, the cGAN model with a convolution kernel size of 6x6 exhibited the best performance on the test set, achieving the following metrics: accuracy of 0.9970, precision of 0.9862, recall of 1.0000, F 1 -score of 0.9930, and IoU of 0.9813. Notably, when compared to other similar grape segmentation studies, the Pix2Pix architecture utilized in this work consistently delivered the best results across all the evaluated metrics. Furthermore, the proposed architecture is able to reconstruct images obtained under noisy and poor illumination conditions indicating its great potential under difficult settings.

Requirements traceability is a technique intended to help determine the impact of changes in software design, support their integration, preserve knowledge, and assure the quality and correctness of the overall system. This paper presents an approach that considers traceability of safety requirements in the context of model-driven development of teleoperated services robots. The combination of the model-driven approach with safety requirements traceability makes it possible to construct systems using techniques for automatically identifying, managing, and mitigating risks so that these systems are safe enough to work in a particular environment. To secure the advantages of these mechanisms, we have developed a tool that provides users with traceability reports after applying model transformations. These reports enable developers to determine whether or not all safety requirements have been considered, the impact of changing a safety requirement, and how they are considered both in architectural decisions and code implementations.