ANFIS-Based Crack Sealing in Concrete Pipes Using a Pipe Crack Sealing Manipulator

In this paper, the Adaptive Neuro-Fuzzy Inference System (ANFIS) is applied to address crack repair in pipes. The methodology involves using a pipe crack sealing manipulator (PCSM), a tree-type robot with five branches for concrete pipe repair, capable of navigating through horizontal pipes, vertical pipes and surfaces. Based on the observations and simulation studies revealed that during the repair operation, only the fifth branch is actively involved. However, when the crack trajectory is sufficiently long, more branches can navigate through horizontal pipes and surfaces. In such cases, traditional analytical approaches fall short in calculating the inverse kinematics, necessitating the use of soft computing methods. To tackle this challenge, the application of the ANFIS technique has been explored to compute the inverse kinematics. To make a meaningful comparison between the ANFIS method and the analytical approach, we focus on the behaviour of only three links of branch five. Based on our findings from the fifth branch of the PCSM, it can be inferred that the ANFIS technique holds promise for addressing the inverse kinematics when more than one branch of PCSM navigates simultaneously.