The study of microgravity, a condition in which an object experiences near-zero weight, is a critical area of research with far-reaching implications for various scientific disciplines. Microgravity allows scientists to investigate fundamental physical phenomena influenced by Earth’s gravitational forces, opening up new possibilities in fields such as materials science, fluid dynamics, and biology. However, the complexity and cost of developing and conducting microgravity missions have historically limited the field to well-funded space agencies, universities with dedicated government funding, and large research institutions, creating a significant barrier to entry. This paper presents the MicroGravity Explorer Kit’s (MGX) design, a multifunctional platform for conducting microgravity experiments aboard suborbital rocket flights. The MGX aims to democratize access to microgravity research, making it accessible to high school students, undergraduates, and researchers. To ensure that the tool is versatile across different scenarios, the authors conducted a comprehensive literature review on microgravity experiments, and specific requirements for the MGX were established. The MGX is designed as an open-source platform that supports various experiments, reducing costs and accelerating development. The multipurpose experiment consists of a Jetson Nano computer with multiple sensors, such as inertial sensors, temperature and pressure, and two cameras with up to 4k resolution. The project also presents examples of codes for data acquisition and compression and the ability to process images and run machine learning algorithms to interpret results. The MGX seeks to promote greater participation and innovation in space sciences by simplifying the process and reducing barriers to entry. The design of a platform that can democratize access to space and research related to space sciences has the potential to lead to groundbreaking discoveries and advancements in materials science, fluid dynamics, and biology, with significant practical applications such as more efficient propulsion systems and novel materials with unique properties.

The geometric characterization of cameras requires laboratory methods during construction and field methods during operation, where measuring multiple points in images is essential to ensure redundancy and reliable results. However, conventional characterization is labor-intensive and demands specific data collection. Recently, efforts have shifted towards deep learning-based solutions to replace repetitive manual tasks. Various learning strategies, neural networks, geometric priors, and datasets have been investigated. This paper focuses on analyzing the performance of some of these AI-based camera characterization techniques, highlighting their advantages, limitations, and potential applications to enhance image-based navigation results.

The detection and re-identification of vehicles in dynamic environments, such as highways monitored by a swarm of drones, presents significant challenges, particularly due to the variability of images captured from different angles and under various conditions. This scenario necessitates the development of suitable methods that integrate appropriate computational techniques, such as convolutional neural networks (CNN) to address the diversity of drone captures and improve accuracy in detection and re-identification. In this paper, a solution for vehicle detection and Re-ID is proposed, combining CNN techniques VGG16, VGG19, ResNet50, InceptionV3 and EfficientNetV2L. YOLOv4 was selected for detection, while the DeepSORT algorithm was chosen for tracking. The proposed solution considers the generalization capabilities of these techniques with varied images from different drones in different positions. Two datasets were employed: the first is a public dataset from Mendeley used for method evaluation, while the second consists of images and data collected by a swarm of drones. In the first experiment, the best performing network was ResNet50, with an average accuracy of 55%. In the second experiment, the highest accuracy CNN was VGG19, with 91% accuracy. Overall, the techniques were able to distinguish vehicles of different models and adapted to the data captured by drones.

This work presents the design, simulation, and validation of an embedded guidance, navigation, and control (GNC) system on a simulated onboard platform for the first stage of a microlauncher vehicle. The objective is to evaluate the performance and feasibility of an embedded control architecture throughout powered flight. The control strategy adopted is a proportional-integral-derivative (PID) controller with gain scheduling, employing gains that are tuned by a linear quadratic (LQ) approach to counteract vehicle dynamics changes. The guidance unit supplies reference attitude angles, and the navigation unit estimates vehicle orientation based on inertial measurements in rotation matrices and Euler angles. The validation process involves software-in-the-loop (SIL) and hardware-in-the-loop (HIL) simulations, with the GNC algorithms executed in real time on an embedded hardware platform. Results show excellent pitch stabilization, accurate tracking of the reference trajectory, and actuator commands within operational limits. The simulated vehicle response is well correlated with expected mission profiles. In addition, functional requirements and verification procedures for the GNC system are also formally stated. These findings justify the development of robust embedded avionics systems for microlaunchers and provide an experimentally verified methodological framework to be applied in the future in small satellite launch vehicles.

Automated surface defect detection in hot-rolled steel is a prerequisite for real-time quality control, yet most deployed inspection systems operate in isolation from Industrial Internet of Things (IIoT) infrastructure. This paper reports a laboratory-scale proof of concept with two contributions. The first is a segmentation study on the Severstal dataset using a leakage-free, defect-stratified split of 1886 test images. Because a trivial all-background predictor already attains 96.66% pixel accuracy, performance is reported through Dice, IoU, precision, recall, and F1 with 95% confidence intervals. A compact from-scratch U-Net (0.49 M parameters) reaches a Dice of 0.416 at 38.6 ms per image, an ImageNet-pretrained DeepLabV3+ model reaches 0.677 at 46.3 ms and 37 times the parameters, and a classical Otsu baseline reaches 0.060, bracketing an explicit accuracy-versus-footprint design space rather than a single recommended model. The second contribution is architectural: a three-layer IIoT architecture whose messaging layer is empirically characterized on a Raspberry Pi broker over 158,500 messages. A factorial experiment isolates the transport configuration of the broker, rather than that of the publisher, as the determinant of end-to-end latency, yielding a seventeen-fold reduction. The layer sustains 1920 messages per second without loss, and a deliberate broker outage shows that MQTT delivery guarantees are semantic rather than temporal, motivating an application-level message-expiry policy. Embedded inference deployment is identified as the primary next step.

The demand for Global Navigation Satellite System (GNSS) positioning processing techniques resilient to low-latitude ionospheric effects, notably signal fading and the related scintillation, has driven the exploration of innovative solutions. Various signal processing and machine learning have been employed to enhance receiver performance under scintillation conditions, but some possible machine learning approaches and simulations still not test. Also, to design and test sophisticated receivers capable of functioning amidst scintillation effects, a dataset of cases of strong scintillation is essential for comprehensive coverage of severe environments. This study introduces a ionospheric amplitude scintillation simulator based on neural networks, utilizing autoencoders, generative adversarial networks (GANs), and particle swarm optimization. The synthetic scintillation simulator generates time series that closely adhere to the statistical characteristics of the α-μ fading model. The proposed simulator comprises a generative adversarial network explicitly trained through supervised learning, preserving the temporal dynamics of the dataset. In addition to the GAN architecture, the simulator includes an autoencoder that learns a low-dimensional latent space, facilitating the reproduction of the temporal relationships observed in historical data by the generator. The implemented simulator was trained and validated using fading and scintillation data collected in São José dos Campos, Brazil during the period of November 16-30, 2014, when scintillation was severe. Results demonstrate that the proposed simulator accurately generates time series with precise values of amplitude scintillation, quantified by the index S4 and first order statistics following the α-μ fading model. This represents the first instance of a simulator achieving such a high degree of statistical fidelity and successful validation, indicating the promising potential of this approach in simulating ionospheric fading channels.