At the core of the system is a Raspberry Pi 5, which serves as the central embedded controller responsible for coordinating every subsystem of the robot. The Raspberry Pi manages motor control, camera operation, sensor processing, wireless communication, odometry calculations, and artificial intelligence inference while maintaining seamless communication between hardware and software modules. Motor speed and direction are controlled using BTS7960 motor drivers, while power distribution is carefully designed to ensure stable operation of all electronic components. This centralized architecture simplifies system integration and enables reliable real-time performance throughout the inspection process.

For visual inspection, the robot is equipped with a Raspberry Pi Camera Module V3, providing continuous live video of the pipeline interior. Instead of relying solely on manual observation, the system incorporates an AI-powered crack detection module based on the YOLOv8 Nano deep learning model. A complete computer vision workflow was implemented, including dataset collection, image annotation, model training, validation, optimization, and deployment directly on the Raspberry Pi 5. The lightweight architecture of YOLOv8 Nano enables real-time inference on embedded hardware while maintaining fast processing speeds and reliable detection performance. Whenever a surface crack is identified, the model highlights the defect and generates a confidence score, allowing operators to quickly identify damaged regions with significantly reduced manual effort. This intelligent inspection approach improves consistency, minimizes human error, and demonstrates the practical application of embedded AI in industrial automation.

In addition to visual inspection, the robot integrates an MQ-4 methane gas sensor to continuously monitor hazardous gas concentrations within pipelines. Since industrial pipelines often transport combustible or toxic gases, real-time environmental monitoring is an essential safety feature. The gas sensing module enables the robot to detect methane leakage during inspection and immediately report sensor readings to the operator. By combining gas detection with computer vision and localization, the robot provides a comprehensive inspection solution capable of monitoring both structural defects and environmental conditions simultaneously.

To provide an intuitive user experience, a browser-based Human Machine Interface (HMI) was developed using the Flask framework. The interface enables wireless communication between the operator and the robot over a Wi-Fi network, eliminating the need for specialized control software. Through this dashboard, users can remotely drive the robot, monitor live camera footage, adjust motor speed, control camera positioning, observe odometry information, and monitor gas sensor readings in real time. The integration of embedded software, networking, and web technologies resulted in a responsive and user-friendly control system that supports efficient remote pipeline inspection.

Throughout the development process, extensive work was carried out across multiple engineering disciplines, including mechanical design, CAD modelling, embedded programming, electronics integration, AI model development, computer vision, web application development, system testing, and performance evaluation. The project required designing the mechanical chassis, selecting suitable hardware components, implementing encoder-based localization algorithms, integrating sensors, developing control software, deploying artificial intelligence models, and validating the complete system under simulated pipeline conditions. This multidisciplinary approach provided practical experience in robotics system integration and reinforced the importance of combining hardware and software into a reliable engineering solution.

Experimental testing demonstrated that the developed prototype successfully achieved stable locomotion inside pipelines, responsive wireless control, continuous live video streaming, reliable encoder-based localization, real-time methane gas monitoring, and embedded AI-based crack detection. The project successfully integrated robotics, artificial intelligence, computer vision, embedded systems, and industrial automation into a unified inspection platform while maintaining a significantly lower implementation cost compared to conventional commercial inspection robots. The resulting system provides a strong foundation for future research in autonomous pipeline inspection, intelligent navigation, IoT integration, cloud-based monitoring, and advanced industrial robotics, making it a scalable solution for modern pipeline maintenance applications.