Department of Electrical, Electronic and Communication Engineering (EECE)

AI-enhanced Digital Twin of Solar Power Plant [ongoing]
This ongoing project is focused on the development of a comprehensive Scalable Digital Twin (DT) model for a large-scale solar photovoltaic (PV) plant. The DT provides real-time monitoring, performance optimization, and predictive maintenance capabilities. As the Co-Principal Invetigator (Co-PI) , my contributions covered end-to-end system design, hardware-software integration, and AI-based predictive modeling. Key technical highlights include:

• Funding: 134,000 USD (Government of Bangladesh Energy Commission).
• it includes developing machine learning models in Python to predict performance degradation and optimize plant efficiency based on historical and real-time data.
• Cloud Integration: It denotes secure and scalable cloud architecture for continuous data logging and visualization dashboards accessible for operators and researchers.
• Cost-Saving Impact: Analyze potential annual savings through predictive maintenance and optimized energy dispatch, reducing reliance on reserve gas and coal-based plants.
• This project aligns with my research interests in Digital Twin Technology and its application in Smart Grids, and AI-based Energy Systems, contributing towards my academic pursuit in the domain of AI-driven intelligent systems.

AI-Driven Smart Conveyor System for Industries [ongoing]

• Funding: 10,000 USD awarded for initial development.
• Funding Source: Military /institute o Science and Technology.
• Objective: To automate product sorting and control conveyor movement using AI-driven machine vision and sensor interfaces.
• Key Features:
  • Telescopic Conveyor Movement: Automated adjustment to product flow and properties.
  • Machine Vision-based Detection: AI-powered detection for product properties like faults, size, and color.
  • Property-based Control: Conveyor movement is adjusted based on product detection results.
• System Components:
  • Programmable Logic Controller (PLC): Core system controller for automation.
  • Optical and Thermal Camera:To provide visual data to the AI-based video processing CPU.
  • Proximity Sensors:
    • Inductive, Capacitive, and Optical sensors to detect diverse product materials and orientations.
  • Actuators:
    • Pneumatic and Electrical Actuators to manage conveyor adjustments.
    Air Compressor: Powers pneumatic systems for smooth operation.
  • HMI Display: Human-machine interface for real-time monitoring and system adjustments.
• Expected Outcomes:
  • Enhanced efficiency in product sorting and defect detection.
  • Minimized human intervention, increasing reliability and speed in material handling.

Smart Street Light Management System for Rural Area

In 2018, I led a team of MIST students in developing an innovative Smart Street Light System tailored for rural environments. Below are the key highlights of the project:

• Funding: $2,000 awarded for prototype development.
• Automatic On/Off Scheduling: Enables efficient energy usage by automating street light operations based on time schedules.
• GSM-SMS Control Mechanism: Allows remote activation, providing flexibility and immediate response to changing lighting needs.
• Implementation: Final system deployed in Mirpur Cantonment, Dhaka.
• Recognition:
  1. Acknowledged by Dhaka City Corporation.
  2. Received the Best Project Award in the MIST project competition.
• Publication: Project was published in IEEE, showcasing a scalable solution for rural infrastructure. Access IEEE Publication.
• Research Alignment: Integrates my research interests in energy-efficient systems and embedded solutions for smart city applications.

AI-Powered Terrestrial Surveillance System
In today’s security landscape, where the need for rapid, accurate threat detection is paramount, AI-driven surveillance systems are crucial for safeguarding critical infrastructure. This project addresses these modern security challenges by providing real-time monitoring and automated response capabilities, tailored specifically for high-priority military and government installations. Leveraging advanced machine learning, the system provides real-time detection and rapid response capabilities, ensuring comprehensive threat monitoring and mitigation. Below are the essential features of this innovative security solution:

• Automated Threat Detection: Capable of identifying fire, weapons, military and civilian vehicles, soldiers, and other potential threats in real time.
• Multi-Channel Alert System:
  • Custom-built radio transceiver circuit and radio handset for direct alert transmission.
  • Alerts routed through GSM network (SMS, voice call), Email, and traditional sirens for multi-tier response.
• Military-Grade Encryption: All communication signals are encrypted to meet high security standards.
• Government Funding: Project was funded by the Bangladesh Military Academy.
• Status:The project is partially published in an Elsevier paper available here.

Based on the success of the developed terretrial surveillance system, another project of Aerial Surveillance is being planned to develop. Custom UAVs will be deployed to apply the experience gathered through the developed project.

Numerous Microcontroller and sensor-based hobbyist projects

• Raspberry-based AI-driven fire detection
• PID-powered seamless line follower
• Smart home security system
• Cost-effective RFID and Fingerprint secured entry system
• Weather-sensing Irrigation System for Agricultural Lands
• Raspberry Pi Controlled AI (YOLO)-driven Fire Detection and Alarm System
• IoT-based Home Automation etc.