Presenting a Distributed Control Method for Guiding a Robotic Swarm in Gas Pipelines for Leak Detection

Given the importance of continuous monitoring of gas transmission pipelines and the need for rapid leak detection, the use of intelligent robotic systems has attracted considerable attention as an innovative approach to energy infrastructure inspection. This study aims to present an optimal distributed control method for guiding a robotic swarm inside gas pipelines in order to detect leaks and ensure safe navigation through complex environments. In the proposed method, the robots’ motion path is modeled as a curved virtual tube, and the D* Lite path-planning algorithm is employed to plan routes in the presence of static and dynamic obstacles. In addition, a deep-learning-based adaptive PID controller is designed to improve guidance performance, reduce tracking error, and enhance system stability. The stability of the control system is evaluated using Lyapunov functions and Chebyshev polynomials. The simulations are conducted in a Python environment, and the performance of the proposed method is examined under various scenarios, including the presence of moving obstacles, complex paths, and multiple leak points. The results demonstrate that the proposed method can accurately follow the optimal path while maintaining robotic swarm cohesion and preventing inter-robot collisions. Furthermore, significant improvements are observed in terms of the time required to reach the target, path-tracking error, and the number of collisions with obstacles compared with benchmark methods. The findings indicate that integrating distributed control, D* Lite path planning, and deep learning can provide an effective solution for the intelligent inspection of gas pipelines and the rapid detection of leaks in complex environments.