Water, one of the most valuable resources, is underutilized in irrigated rice production. The yield of rice, a staple food across the world, is highly dependent on having proper irrigation systems. Alternate wetting and drying (AWD) is an effective irrigation method mainly used for irrigated rice production. However, unattended, manual, small-scale, and discrete implementations cannot achieve the maximum benefit of AWD. Automation of large-scale (over 1000 acres) implementation of AWD can be carried out using wide-area wireless sensor network (WSN). An automated AWD system requires three different WSNs: one for water level and environmental monitoring, one for monitoring of the irrigation system, and another for controlling the irrigation system. Integration of these three different WSNs requires proper dimensioning of the AWD edge elements (sensor and actuator nodes) to reduce the deployment cost and make it scalable. Besides field-level monitoring, the integration of external control parameters, such as real-time weather forecasts, plant physiological data, and input from farmers, can further enhance the performance of the automated AWD system. Internet of Things (IoT) can be used to interface the WSNs with external data sources. This research focuses on the dimensioning of the AWD system for the multilayer WSN integration and the required algorithms for the closed loop control of the irrigation system using IoT. Implementation of the AWD for 25,000 acres is shown as a possible use case. Plastic pipes are proposed as the means to transport and control proper distribution of water in the field, which significantly helps to reduce conveyance loss. This system utilizes 250 pumps, grouped into 10 clusters, to ensure equal water distribution amongst the users (field owners) in the wide area. The proposed automation algorithm handles the complexity of maintaining proper water pressure throughout the pipe network, scheduling the pump, and controlling the water outlets. Mathematical models are presented for proper dimensioning of the AWD. A low-power and long-range sensor node is developed due to the lack of cellular data coverage in rural areas, and its functionality is tested using an IoT platform for small-scale field trials.

Microvascular complications are one of the key causes of mortality among type 2 diabetic patients. This study was sought to investigate the use of a novel machine learning approach for predicting these complications using only the patient demographic, clinical, and laboratory profiles. A total of 96 Bangladeshi participants with type 2 diabetes were recruited during their routine hospital visits. All patient profiles were assessed by using a chi-squared (χ²) test to statistically determine the most important markers in predicting three microvascular complications: cardiac autonomic neuropathy (CAN), diabetic peripheral neuropathy (DPN), and diabetic retinopathy (RET). A machine learning approach based on logistic regression, random forest (RF), and support vector machine (SVM) algorithms was then developed to ensure automated clinical testing for microvascular complications in diabetic patients. The highest prediction accuracies were obtained by RF using diastolic blood pressure, albumin–creatinine ratio, and gender for CAN testing (98.67%); microalbuminuria, smoking history, and hemoglobin A1C for DPN testing (67.78%); and hemoglobin A1C, microalbuminuria, and smoking history for RET testing (84.38%). This study suggests machine learning as a promising automated tool for predicting microvascular complications in diabetic patients using their profiles, which could help prevent those patients from further microvascular complications leading to early death.

The regionally adaptable nutrition-centric food recommendation system (FR-RANC) introduces a holistic approach to personalized dietary recommendations that addresses the limitations of existing nutrition systems that often overlook regional food availability, economic conditions, and comprehensive nutrient requirements. Utilizing a decision support system (DSS) and an optimization algorithm, the system integrates user-specific data, including age, gender, health conditions, and socioeconomic status, with region-specific food databases to generate food lists that meet macro-and micronutrient needs. Unlike traditional systems, which focus on a few nutrients or specific medical conditions, FR-RANC provides a holistic solution by considering a broader range of nutrients and adapting them to various regional and economic contexts. The initial evaluation, based on systematically generated test cases, demonstrated the system’s capability to achieve approximately 90% compliance with macronutrient targets. At the same time, micronutrient results exhibit a narrow distribution around the mean, indicating strong alignment with recommended requirements, even across varying health conditions, such as diabetes, cardiovascular issues, and pregnancy. Furthermore, FR-RANC effectively addresses cold-start and scalability issues, as observed in traditional systems. These results underscore its potential as a proactive tool for promoting balanced nutrition, supporting healthcare professionals, and addressing global nutritional disparities.

Major Depressive Disorder (MDD) is a serious mental disorder that if untreated not only affects physical health but also has a high risk of suicide. While the neurophysiological phenomena that contribute to the formation of Suicidal Ideation (SI) are still ill-defined, clear links between MDD and cardiovascular disease have been reported. The aim of this study is to extract suitable features from arterial pulse signals with a view to predicting SI within MDD and control groups. Sixteen unmedicated MDD patients with a history of SI (MDDSI+), sixteen without SI (MDDSI-) and twenty-nine healthy subjects (CONT) were recruited at a psychiatric clinic in the UAE. Depression severity and SI were assessed using the Hamilton Depression Rating Scale and Beck Depression Inventory. Pulse Wave Amplitude (PWA) was calculated as the difference between the peak (Systole) and the valley (Diastole) of the arterial pulse within each cardiac cycle. Then, 2D Tone-Entropy (TE) features were extracted from the Systole, Diastole and PWA time series. The TE features extracted from Diastole were the best markers for predicting MDDSI+. The overall classification accuracies of Classification and Regression Tree (CART) model by using TE features of Systole, Diastole and PWA were 88.52%, 90.2% and 88.52% respectively. When all TE features were combined, accuracy increased up to 93.44% in identifying MDDSI+/MDDSI-/Control groups.

Internet of Things (IoT) and wireless sensor networks (WSN) have significant potential for a wide range of applications. However, these methods are hindered by resource constraints, long-range connectivity and energy challenges. Existing clustering protocols attempt to address some of these problems but fall short when handling the demands of real-time mobility in large-scale WSNs. Many of these algorithms rely on high processing power and offline optimization and fail to account for the dynamic mobility of nodes within the network. This article introduces the Multi-hop Mobility-based Real-Time Clustering (MHM-RTC) algorithm, which utilizes long-range (LoRa) communication between mobile sensor nodes (SNs) and a mobile data sink (DS). Unlike traditional approaches, this algorithm enables dynamic, on-the-fly clustering through the active participation of both SNs and DS. It adapts to continuous topology changes in highly mobile environments, supporting real-time responsiveness and efficient wide-area WSN coverage. The algorithm includes a mathematical model to calculate the clustering time across various network sizes, focusing on energy optimization and extended coverage in mobile networks. The simulation results demonstrate significant improvements over existing mobility-supporting algorithms, such as the Energy Efficient Routing Algorithm (EERA), Energy-Efficient Mobility-Based Cluster Head Selection (EEMCS) and Lightweight Dynamic Clustering Algorithm (LDCA). Specifically, MHM-RTC reduced the clustering time by up to 20%, increased the sensor node coverage by 37-53%, and reduced the average residual energy consumption by 3.77–8.69% when the total number of sensor nodes varied from 100 to 250. These results highlight that MHM-RTC is an effective solution for real-time energy-efficient clustering in large-scale mobile sensor networks.

Microvascular complications are one of the key causes of mortality among type-2 diabetic patients. This study was sought to investigate the use of a novel machine learning approach for predicting these complications from patient demographic, clinical, and laboratory profiles only. A total of 96 Bangladeshi participants having type-2 diabetes were recruited during their routine hospital visits. All patient profiles were assessed by using a Chi-squared (2) test to statistically determine the most important markers in predicting four microvascular complications; namely cardiac autonomic neuropathy (CAN), diabetic peripheral neuropathy (DPN), diabetic nephropathy (NEP), and diabetic retinopathy (RET). A machine learning approach based on random forest (RF) and support vector machine (SVM) was then developed to ensure automated clinical testing for microvascular complication in diabetic patients. The highest prediction accuracies were obtained by RF using diastolic blood pressure, Albumin-Creatinine ratio, and gender for CAN testing (98.67%), Microalbuminuria, smoking history, and hemoglobin A1C for DPN testing (67.78%), Albumin-Creatinine ratio for NEP testing (100%), and hemoglobin A1C, Microalbuminuria, and smoking history for RET testing (84.38%). This study suggests machine learning as a promising automated tool for predicting microvascular complications in diabetic patients using their profiles, which could help prvent those patients from further microvascular complications leading to early death.