Deep Learning models have presented promising results when applied to Agriculture 4.0. Among other applications, these models can be used in disease detection and fruit counting. Deep Learning models usually have many layers in the architecture and millions of parameters. This aspect hinders the use of Deep Learning on mobile devices as they require a large amount of processing power for inference. In addition, the lack of high-quality Internet connectivity in the field impedes the usage of cloud computing, pushing the processing towards edge devices. This work describes the proposal of an edge AI application to detect and map diseases in citrus orchards. The proposed system has low computational demand, enabling the use of low-footprint models for both detection and classification tasks. We initially compared AI algorithms to detect fruits on trees. Specifically, we analyzed and compared YOLO and Faster R-CNN. Then, we studied lean AI models to perform the classification task. In this context, we tested and compared the performance of MobileNetV2, EfficientNetV2-B0, and NASNet-Mobile. In the detection task, YOLO and Faster R-CNN had similar AI performance metrics, but YOLO was significantly faster. In the image classification task, MobileNetMobileV2 and EfficientNetV2-B0 obtained an accuracy of 100%, while NASNet-Mobile had a 98% performance. As for the timing performance, MobileNetV2 and EfficientNetV2-B0 were the best candidates, while NASNet-Mobile was significantly worse. Furthermore, MobileNetV2 had a 10% better performance than EfficientNetV2-B0. Finally, we provide a method to evaluate the results from these algorithms towards describing the disease spread using statistical parametric models and a genetic algorithm to perform the parameters’ regression. With these results, we validated the proposed pipeline, enabling the usage of adequate AI models to develop a mobile edge AI solution.

Ecological environments research helps to assess the impacts on forests and managing forests. The usage of novel software and hardware technologies enforces the solution of tasks related to this problem. In addition, the lack of connectivity for large data throughput raises the demand for edge-computing-based solutions towards this goal. Therefore, in this work, we evaluate the opportunity of using a Wearable edge AI concept in a forest environment. For this matter, we propose a new approach to the hardware/software co-design process. We also address the possibility of creating wearable edge AI, where the wireless personal and body area networks are platforms for building applications using edge AI. Finally, we evaluate a case study to test the possibility of performing an edge AI task in a wearable-based environment. Thus, in this work, we evaluate the system to achieve the desired task, the hardware resource and performance, and the network latency associated with each part of the process. Through this work, we validated both the design pattern review and case study. In the case study, the developed algorithms could classify diseased leaves with a circa 90% accuracy with the proposed technique in the field. This results can be reviewed in the laboratory with more modern models that reached up to 96% global accuracy. The system could also perform the desired tasks with a quality factor of 0.95, considering the usage of three devices. Finally, it detected a disease epicenter with an offset of circa 0.5 m in a 6 m × 6 m × 12 m space. These results enforce the usage of the proposed methods in the targeted environment and the proposed changes in the co-design pattern.

O crescente interesse em tecnologias de Realidade Aumentada (RA) está aumentando devido ao seu uso recente por diversos setores produtivos da indústria. Também, por Indústria 4.0, um dos pilares é RA e inteligência artificial (IA) para apoiar colaboração homem-máquina. A IA centrada no ser humano é uma pesquisa emergente área com foco no fornecimento de soluções interativas para apoiar as atividades humanas. Este artigo apresenta um aplicativo móvel (App) inovador composto por um RA interface conectada a uma IA para auxiliar um usuário na tomada de decisão interativa em a inspeção laranja.

Human Activity Recognition (HAR) is a topic of interest in several areas, for example, health and sports. There are several ways to perform HAR tasks, which must be understood before creating the desired tools. In this article, we carry out a theoretical study to understand the HAR problem and its main techniques using AI. In addition, we present a case study in which we developed a prototype HAR system for walking and trekking. In this study, we evaluated new hardware for HAR with IMUs with 9 degrees of freedom. This system, composed of four sensors called SPUs was attached to the users lower limbs. A device called WPU collects information from these sensors. With the fusion of the collected data, AI techniques were applied with CNN models reaching an accuracy of 98% to classify the human activities in this context, thus creating new perspectives for using AI to HAR in sports.

In this century, smart devices are increasingly present in our lives, such as at work, sports, or household chores. In this context, we have wearable devices that can help people with health monitoring or physical performance in sports activities. With the integration of artificial intelligence (AI), these wearable devices can identify injuries in athletes or care for the elderly in rehabilitation from human activity recognition (HAR). AI techniques are commonly applied for pattern recognition, such as image classification or HAR. In this context, we seek to develop a smart wearable device to recognize walking activities. In order to improve the identification of these tasks through AI algorithms, we propose the fusion of data between four sensors called SPUs. Each SPU has NodeMCU ESP-32 and BNO080 IMU hardware in its architecture. The data from these hardware provides information in high precision. A zero W raspberry pi collected this information. After extracting and manipulating this data, we trained a deep learning model. The model accuracy was higher than 92% reaching an overall accuracy of 97%. Therefore, the smart wearable device showed a new tool for recognizing walking activity, which could be applied in the future to recognize more complex tasks.

Simulating real-world activities in virtual environments with high fidelity is crucial for advancing training, simulation, and data analysis. This capability is particularly advantageous for complex, hazardous, or impractical tasks in real settings, as it enables extensive investigation without compromising safety or feasibility. This paper introduces a system for capturing and virtually simulating walking and running motions by utilizing gyroscope data to monitor the rotational dynamics of the lower limbs. Our approach provides an accurate recording of these movements, supporting a detailed biomechanical analysis. Our findings demonstrate that specialized sensors, as opposed to general-purpose devices like smartwatches, yield more precise, targeted data, thereby enhancing the quality of movement analysis.We also consider the potential of integrating multimodal data, such as combining motion sensors with image-based analysis for enhanced human activity recognition. Although not yet implemented, this integration offers a promising direction for future system improvements, aiming to increase the scope and precision of activity recognition in virtual environments. Our system lays foundational work for advanced human data capture devices, creating a link between the physical and virtual realms. This enables comprehensive studies of biomechanics with potential applications in fields like sports science and orthopedics.