The consumption of specialty coffee has increased around the world. Specialty coffee is free from impurities and defects. Specialty coffee is produced in smaller quantities, and its production process is hard and expensive. Traditional coffee beans have defects that affect the flavor of the coffee. It is essential to select the type of coffee to guarantee the best cost and quality. When a human manually classifies the type of coffee, there may be interference due to the human condition. This process has the disadvantage of being subjective. Few studies used Machine Learning methods for predicting specialty coffee classification by analyzing images. This article proposes a framework for classifying specialty coffees by applying Tiny Machine Learning techniques. We developed a model that can help accurately analyze and classify the coffee process and inform the quality of coffee without human interference. We achieved 100% accuracy, and we believe our model can be used effectively in the coffee industry.

The digital transformation in healthcare is a recurring theme around the world. However, the continental extension of Brazil makes it difficult to offer basic healthcare services in remote locations. Due to the scarcity of diagnosis services in these areas, this article proposes leveraging Federated Learning as a way to reduce costs and mitigate these problems, helping with pre-diagnosis. The experiments carried out showed that just using the federated approach can increase the model's predictive capacity and reduce training time. The model developed using Federated Learning increased the model's accuracy by 14%, while managing to reduce the model's loss in the validation set by 1.0354.

In the realm of agribusiness, transformative shifts are underway, propelled by the growing demands and expanding scales of grain production. This evolution calls for a critical reevaluation of the existing paradigms in coffee production and marketing paradigms, with a specific focus on integrating Artificial Intelligence (AI). This work aims to review, synthesize, and summarize the available data regarding how Machine Learning (ML) has been used to detect and classify characteristics in coffee beans and leaves. For this purpose, a comprehensive literature review of the most significant research contributions describing the application of AI for advanced classification techniques in coffee agriculture has been carried out. Our analysis suggests that implementing AI technologies allows the classification of coffee, encompassing various attributes such as maturity, roast intensity, disease identification, flavor profiles, and overall quality. More largely, this technological advancement holds the potential to revolutionize coffee farming by providing producers and agricultural specialists with sophisticated tools to enhance production efficiency, minimize costs, and improve the accuracy and confidence of their decision-making processes.

Reconfigurable Intelligent Surfaces (RIS) are a class of metamaterials that have gained significant attention in recent years due to their potential to revolutionize wireless communication, sensing, and imaging technologies. RISs consist of a planar array of closely spaced, subwavelength-sized elements that can manipulate electromagnetic waves in a controllable manner. By reconfiguring the geometry, material properties, or phase of individual elements on the RIS, the surface can be customized to meet specific application requirements. RISs can potentially improve wireless communication by creating virtual channels, reducing interference, and improving overall quality. They can also enhance the efficiency of energy harvesting systems and improve sensing and imaging technologies by manipulating the propagation and scattering of electromagnetic waves. Additionally, RIS could be used to increase privacy and security by selectively blocking or allowing certain frequencies of electromagnetic waves. In this article, we provide a brief history of the development of RIS and discuss the design and fabrication of RIS structures. We also explore the potential applications and benefits of RIS technology, including improved wireless communication, enhanced energy efficiency, advanced sensing and imaging, and increased privacy and security. Finally, we highlight some of the current research challenges and future directions for RIS technology. Overall, RISs hold great promise for advancing a wide range of technologies and applications, and we expect to see many exciting developments in this area in the future.

It is known that the Central Limit Theorem (CLT) is not always the most appropriate tool for deriving closed-form expressions. We evaluate a Single-Input Single-Output (SISO) system performance in which the Large Intelligent Surface (LIS) acts as a scatterer. The direct link between the transmitting and receiving devices is negligible. Quantization phase errors are considered since the high precision configuration of the reflection phases is not always feasible. We derive exact closed-form expressions for the spectral efficiencies, outage probabilities, and average symbol error rate (SER) of different modulations. We assume a more comprehensive scenario in which $b$ bits are dedicated to the LIS elements’ phase adjustment. From Monte Carlo simulations, we prove the excellent accuracy of our approach and investigate the behavior of power scaling law and power required to reach a specific capacity, depending on the number of reflecting elements. We show that the LIS with approximately fifty elements and four dedicated bits for phase quantization outperforms the conventional system performance without LIS.

In this paper, we use a series expansion to calculate the sum-capacity of a massive Multiple-Input MultipleOutput (MIMO) system under propagation environment described by a dominant line-of-sight. The sum-capacity is written as Taylor’s series where each term is a function of the mean trace of k-th power of the channel matrix W. We analytically derive the mean trace of first, second, third, and fourth moments of W. Although the series is infinite, our numerical results show that only a few terms can tightly approximate the exact sum capacity. Numerical results corroborate our analytical expressions.

In this letter, we present a study on linear channel estimators and their respective mean square error (MSE) expressions acknowledging spatially correlated channels and pilot contamination. We also investigate the impact of imperfect channel covariance matrix knowledge.

The use of large-scale antenna arrays grants considerable benefits in energy and spectral efficiency to wireless systems due to spatial resolution and array gain techniques. By assuming a dominant line-of-sight environment in a massive MIMO scenario, we derive analytical expressions for the sum-capacity. % Then, we show that convenient simplifications on the sum-capacity expressions are possible when working at low and high SNR regimes. % Furthermore, in the case of a high SNR regime, it is demonstrated that the Gamma PDF can approximate the PDF of the instantaneous channel sum-capacity as the number of BS antennas grows. A second important demonstration presented in this work is that a Gamma PDF can also be used to approximate the PDF of the summation of the channel’s singular values as the number of devices increases. Finally, it is important to highlight that the presented framework is useful for a massive number of Internet of Things devices as we show that the transmit power of each device can be made inversely proportional to the number of BS antennas.