MmWave Physical Layer Network Modeling and Planning for Fixed Wireless Access Applications.

The large bandwidths that are available at millimeter-wave frequencies enable fixed wireless access (FWA) applications, in which fixed point-to-point wireless links are used to provide internet connectivity. In FWA networks, a wireless mesh is created and data are routed from the customer premises equipment (CPE) towards the point of presence (POP), which is the interface with the wired internet infrastructure. The performance of the wireless links depends on the radio propagation characteristics, as well as the wireless technology that is used. The radio propagation characteristics depend on the environment and on the considered frequency. In this work, we analyzed the network characteristics of FWA networks using radio propagation models for different wireless technologies using millimeter-wave (mmWave) frequencies of 28 GHz, 60 GHz, and 140 GHz. Different scenarios and environments were considered, and the influence of rain, vegetation, and the number of subscribers was investigated. A network planning algorithm is presented that defines a route for each CPE towards the POP based on a predefined location of customer devices and considering the available capacity of the wireless links. Rain does not have a considerable effect on the system capacity. Even though the higher frequencies exhibit a larger path loss, resulting in a lower power of the received signal, the larger bandwidths enable a higher channel capacity.

Evaluating 60 GHz FWA Deployments for Urban and Rural Environments in Belgium.

Fixed wireless access (FWA) provides a solution to compete with fiber deployment while offering reduced costs by using the mmWave bands, including the unlicensed 60 GHz one. This paper evaluates the deployment of FWA networks in the 60 GHz band in realistic urban and rural environment in Belgium. We developed a network planning tool that includes novel backhaul based on the IEEE 802.11ay standard with multi-objective capabilities to maximise the user coverage, providing at least 1 Gbps of bit rate while minimising the required network infrastructure. We evaluate diverse serving node locations, called edge nodes (EN), and the impact of environmental factors such as rain and vegetation on the network design. Extensive simulation results show that defining a proper EN's location is essential to achieve viable user coverage higher than 95%, particularly in urban scenarios where street canyons affect propagation. Rural scenarios require nearly 75 ENs per km2 while urban scenarios require four times (300 ENs per km2) this infrastructure. Finally, vegetation can reduce the coverage by 3% or increment infrastructure up to 7%, while heavy rain can reduce coverage by 5% or increment infrastructure by 15%, depending on the node deployment strategy implemented.

Prediction of patient admission and readmission in adults from a Colombian cohort with bipolar disorder using artificial intelligence.

Introduction: Bipolar disorder (BD) is a chronically progressive mental condition, associated with a reduced quality of life and greater disability. Patient admissions are preventable events with a considerable impact on global functioning and social adjustment. While machine learning (ML) approaches have proven prediction ability in other diseases, little is known about their utility to predict patient admissions in this pathology. Aim: To develop prediction models for hospital admission/readmission within 5 years of diagnosis in patients with BD using ML techniques. Methods: The study utilized data from patients diagnosed with BD in a major healthcare organization in Colombia. Candidate predictors were selected from Electronic Health Records (EHRs) and included sociodemographic and clinical variables. ML algorithms, including Decision Trees, Random Forests, Logistic Regressions, and Support Vector Machines, were used to predict patient admission or readmission. Survival models, including a penalized Cox Model and Random Survival Forest, were used to predict time to admission and first readmission. Model performance was evaluated using accuracy, precision, recall, F1 score, area under the receiver operating characteristic curve (AUC) and concordance index. Results: The admission dataset included 2,726 BD patients, with 354 admissions, while the readmission dataset included 352 patients, with almost half being readmitted. The best-performing model for predicting admission was the Random Forest, with an accuracy score of 0.951 and an AUC of 0.98. The variables with the greatest predictive power in the Recursive Feature Elimination (RFE) importance analysis were the number of psychiatric emergency visits, the number of outpatient follow-up appointments and age. Survival models showed similar results, with the Random Survival Forest performing best, achieving an AUC of 0.95. However, the prediction models for patient readmission had poorer performance, with the Random Forest model being again the best performer but with an AUC below 0.70. Conclusion: ML models, particularly the Random Forest model, outperformed traditional statistical techniques for admission prediction. However, readmission prediction models had poorer performance. This study demonstrates the potential of ML techniques in improving prediction accuracy for BD patient admissions.

Performance Evaluation of Direct-Link Backhaul for UAV-Aided Emergency Networks.

Today's wireless networks provide us reliable connectivity. However, if a disaster occurs, the whole network could be out of service and people cannot communicate. Using a fast deployable temporally network by mounting small cell base stations on unmanned aerial vehicles (UAVs) could solve the problem. Yet, this raises several challenges. We propose a capacity-deployment tool to design the backhaul network for UAV-aided networks and to evaluate the performance of the backhaul network in a realistic scenario in the city center of Ghent, Belgium. This tool assigns simultaneously resources to the ground users-access network-and to the backhaul network, taking into consideration backhaul capacity and power restrictions. We compare three types of backhaul scenarios using a 3.5 GHz link, 3.5 GHz with carrier aggregation (CA) and the 60 GHz band, considering three different types of drones. The results showed that an optimal UAV flight height (80 m) could satisfy both access and backhaul networks; however, full coverage was difficult to achieve. Finally, we discuss the influence of the flight height and the number of requesting users concerning the network performance and propose an optimal configuration and new mechanisms to improve the network capacity, based on realistic restrictions.

IMPORTANCIA DE LAS FRECUENCIAS DE RESONANCIA DEL TRACTO VOCAL EN LA ESTIMACIÓN DE POSICIONES ARTICULATORIAS / IMPORTANCE OF THE RESONANCE FREQUENCIES OF THE VOCAL TRACT IN ESTIMATING ARTICULATORY POSITIONS

La inversión articulatoria, cuyo objetivo es estimar la posición de los órganos articuladores a partir de la información contenida en la señal de voz, ofrece una variedad de potenciales aplicaciones en el campo de la voz; sin embargo, este es un problema aún por resolver. En este sentido, buscar representaciones con la capacidad de incrementar el desempeño de los sistemas de inversión articulatoria es una tarea importante. El presente trabajo analiza la relevancia de los formantes como entrada para los sistemas de inversión articulatoria. Para ello se implementa un análisis analítico y estadístico. En el caso analítico se utiliza un sintetizador articulario, el cual simula la ecuación de tubos concatenados que modelan el tracto vocal. Para el análisis estadístico se estudian datos reales provenientes de un articulógrafo electromagnético para los cuales se estima la asociación entre las características acústicas y los movimientos de los órganos articuladores. A modo de medida de asociación estadística se utiliza la medida de información . Los resultados entregados por el análisis son corroborados en un sistema de inversión articulatoria basado en redes neuronales. Se observa una mejora en el valor de error cuadrático medio del 2,2% y para el caso de la medida de desempeño de la correlación, una mejora del 2,8%.

Acoustic-to-Articulatory inversion, which seeks to estimate an articulator position using the acoustic information in the speech signal, offers several potential applications in the field of speech processing. In this context, it is important to use acoustic parameters with the ability to increase the performance of acoustic-to-articulatory inversion systems. This paper analyzes the importance of formants as inputs to such inversion systems from an analytical and a statistical perspective. The former is based on an articulatory synthesizer that simulates the voice signal from the vocal tract. The statistical analysis is based on real data provided by an electromagnetic articulograph, for which we estimate the statistical association between acoustic features and articulator movement. As a measure of statistical association, the information measure is utilized. The results are tested on a neuralnetwork- based Acoustic-to-Articulatory inversion system. The use of formants as inputs led to an improvement of 2.2% and 2.8% in the root-mean-square error and correlation values, respectively.

Kernel Principal Component analysis through time for voice disorder classification.

Kernel Principal Component analysis is a nonlinear generalization of the popular linear multivariate analysis method. However, this method assumes that the observed data is independent, a disadvantage for many practical applications. In order to overcome this difficulty, the authors propose a combination of Kernel Principal Component analysis and hidden Markov models. The novelty of the proposed method consists mainly in the way in which a static dimensionality reduction technique has been combined with a classic mixture model in time, to enhance the capabilities of transformation, reduction and classification of voice disorder data. Experimental results show improvements in classification accuracies even with highly reduced representations of the two databases used.

Identification of spike sources using proximity analysis through hidden Markov models.

Hidden Markov models have shown promising results for identification of spike sources in Parkinson's disease treatment, e.g., for deep brain stimulation. Usual classification criteria consist in maximum likelihood rule for the recognition of the correct class. In this paper, we present a different classification scheme based in proximity analysis. For this approach matrices of Markov process are transformed to another space where similarities and differences to other Markov processes are better revealed. The authors present the proximity analysis approach using hidden Markov models for the identification of spike sources (Thalamo and Subthalamo sources, Gpi and GPe sources). Results show that proximity analysis improves recognition performance for about 5% over traditional approach.

Detección en línea de onda-p en sismogramas 3-d utilizando redes neuronales

Se desarrolla un detector en línea del tiempo de arribo de ondas P sobre registros electrónicos de tres componentes de eventos tectónicos. Se lleva a cabo la detección en línea de la onda-p empleando RN, empleando dos diferentes formas: en la primera se sintetiza el detector de LRV, mientras en la segunda se desarrolla un clasificador estadístico con las respectivas clases (señal con onda-p presente y señal con solo ruido presente). La detección se realiza empleando redes neuronales de tipo perceptrón multi-capa, en las cuales se aprovechan las siguientes cualidades: su capacidad de mapeo no lineal de entrada-salida, buena generalización, bajo costo computacional, entre otras. Las anteriores cualidades hacen que las RN sean apropiadas para el proceso de señales en tiempo real. Las entradas a la RN corresponden a datos normalizados del registro sísmico y las características que miden el grado de polarización de la onda. Los registros procesados se obtuvieron de la base de datos del instituto IRIS y de Ingeominas, y su análisis dan como resultado un mejor desempeño del sistema en la detección del tiempo de arribo, aunque no presenta una exactitud aceptable en la estimación del tiempo de arribo de la onda P. (AU)

Detección en línea de onda-p en registros sísmicos utilizando métodos estadísticos

Se presenta el desarrollo de estimadores en línea del tiempo de arribo de ondas-p en registros sísmicos de una componente, correspondientes a sismos tectónicos del tipo regional y local. La derivación formal de algoritmos de detección de cambios abruptos se realiza dentro de la clase de métodos estadísticos de análisis de procesos aleatorios estacionarios, para los cuales se consideran dos modelos de procesos aleatorios: completamente independientes y los basados en modelos regresivos (AR). La prueba de hipótesis empleada corresponde al logaritmo de relación de verosimilitud (LRV), de la cual se derivan los algoritmos analizados: la suma acumulativa (CUSUM), el LRV generalizado (GLRV), 2 CUSUM y CUSUM bilateral. Los registros de entrenamiento y prueba se obtuvieron de la base de datos del instituto IRIS, y de Ingeominas - Manizales. Los resultados del análisis mostraron el buen desempeño en el algoritmo CUSUM normalizado con un error promedio de 21

y una relación señal a ruido mínima promedio de 9.4dB. Sin embargo, el nivel falsas alarmas es alto resultado de la sintonización de los parámetros del estimador que son muy dependientes de cada sismograma. (AU)