International Journal of Artificial Intelligence of Things (AIoT) in Communication Industry

Application of Artificial Neural Network Control Techniques for Torque Ripple Mitigation in Bearingless Synchronous Motors

Baridakara Deesor — 2026-02-16

This paper investigates the application of Artificial Neural Network (ANN)-based control techniques for mitigating torque ripple in bearingless synchronous motors, a critical challenge in high-precision electric drive systems. The increasing deployment of electric motors in industrial automation, robotics, renewable energy, and aerospace applications has intensified the demand for advanced control strategies that ensure smooth torque production, reduced vibration, and reliable long-term operation. Torque ripple in synchronous motor drives leads to periodic torque pulsations that cause mechanical stress, acoustic noise, performance degradation, and reduced service life. The main objective of this study is to design and evaluate an ANN-based control framework capable of effectively minimizing torque ripple under varying operating conditions. The proposed method exploits the nonlinear mapping and adaptive learning capabilities of neural networks to dynamically adjust control parameters in response to system variations. A detailed motor drive model was developed and implemented in the MATLAB/Simulink environment using operational data obtained from industrial installations to enhance practical relevance. Two industrial motor samples from Nigeria were used for validation: an 11.0 kW, 400 V motor from the BUA industrial facility and an 11.0 kW, 690 V motor from the Indorama plant, both with distinct speed and efficiency ratings. Simulation results show that the ANN-based controller achieved approximately 35.4% reduction in torque ripple compared with conventional control methods, leading to smoother torque profiles, reduced mechanical loading, lower noise emission, and improved control accuracy. The findings confirm the effectiveness and robustness of ANN-based control for torque ripple mitigation in bearingless synchronous motor applications.

AgriBot: IoT-enabled Autonomous Robotic System for Real-time Soil and Environmental Monitoring in Precision Agriculture

Abdullateef Ogundipe — 2026-01-02

Agriculture, as a foundation of food security and economic development, increasingly depends on technological innovations to address rising demands for efficiency, sustainability, and productivity. Traditional farming practices often struggle with inefficiencies in monitoring and resource management, creating a need for intelligent solutions. This study presents AgriBot, an autonomous agricultural robot designed to support precision farming through real-time monitoring of soil and environmental conditions. Leveraging IoT-enabled sensors, embedded microcontrollers, and a web-based dashboard, AgriBot collects and analyzes data to provide actionable insights for farmers. Experimental evaluation showed that the proposed system achieved approximately 95% obstacle detection accuracy during autonomous navigation and about 90% accuracy in soil moisture sensing across varying soil conditions. Additionally, real-time control and monitoring were supported with response latencies below 500 ms and continuous operation of up to 4–5 hours, demonstrating the system’s suitability for practical precision agriculture applications.

AI-assisted GEOINT for Border Security: A LEO-centric Architecture Integrating IMINT, UAV Reconnaissance, and OSINT

Settapong Malisuwan — 2026-01-27

Border security operations require persistent monitoring, timely detection, and reliable decision support across large and diverse geographic areas. Traditional approaches based on single sensors or isolated intelligence sources often struggle to provide comprehensive situational awareness under changing environmental and operational conditions. This paper proposes an AI-assisted geospatial intelligence (GEOINT) architecture to support military and security border operations by integrating Low Earth Orbit (LEO) satellite imagery intelligence (IMINT), Unmanned Aerial Vehicle (UAV) reconnaissance, and Open-Source Intelligence (OSINT) within a unified framework. The proposed architecture adopts a layered approach in which LEO satellites provide wide-area and persistent monitoring, optical and synthetic aperture radar (SAR) sensors offer complementary observation capabilities, and UAVs serve as tactical gap-fillers for localized verification and responsive reconnaissance. Artificial intelligence is embedded within the processing layer to support object detection, change detection, and information fusion, with the primary goal of assisting analysts in managing large data volumes and prioritizing areas of interest. Human-in-the-loop control is maintained throughout the workflow to ensure accountability, transparency, and operational relevance. Rather than presenting performance metrics or sensitive implementation details, the paper focuses on architectural design, operational roles, and illustrative border security scenarios. The results highlight how the integration of satellite IMINT, UAV data, OSINT, and AI-assisted analysis can enhance situational awareness and decision support while remaining aligned with established GEOINT doctrine and ethical considerations. The proposed framework provides a practical foundation for future development of border security GEOINT systems.