The Research Of Pipeline Robot Passing Through Pipe Based On Multi-sensor Fusion

Pipeline robot is an intelligent device for human services in the special circumstances, with the development of pipeline inspection technology and application of the maintenance needs of pipeline in service robots for the research and application of pipe provided a broad space for development, this field of research has gotten the world's attention. Because of the pipeline special circumstances, the working narrow, and the need for long-range walking, control of environmental detection and walking research is very necessary. After years of development, the pipeline robot technology has gotten the rapid development , from the overall technical point of view, but research and development of pipeline robot is still in development stage, the results mainly concentrated in the experimental prototype, which still exist some distance from the industrial application .Most of the current limitations of the research pipeline robot in the straight tube environment, intelligence is not very high, walking in the pipeline is not very stable, especially in the tapered tube and corners. For the three- driven robot, passing through elbow pipe and over obstacle have been the bottleneck of the control of the robot walking. In this paper, multi-sensor information fusion algorithm and the PMAC controller are used to control and turn to walk through the nature of research in the pipeline.This paper firstly introduces the research background and current status of the pipe-robot and information fusion algorithm , according to the characteristics of oil and gas pipelines environment, pipeline robot walking mechanism used by three differential drive mode, after the design of such a structure can be achieved only in the pipeline steady walking, but also to achieve cornering and obstacle. This work are summarized as follows:After the introduction of fuzzy neural network based on the established oil and gas pipe robot for fuzzy neural network model to determine the input and output, and fuzzy rules, and had to learn the derivation operator and finally the velocity change in the pipeline was simulated.The arc corners and T-Curve are described ,geometric constraints and motion constraints of the robot in the corners are analyzed, focusing on the robot motion control in the T-Curve, robot collision detection and across obstacles ability when steps and obstacles encountered , the PMAC controller was used to control the robot, and then the robot's speed simulation analysis of front and rear wheel was done when passing passing through elbow pipe.