Rolling Prevention Mechanism for Underground Pipe Erosion Inspection Robot with a Real Time Vision System

Abstract

Pipe inspection is one of the areas that have attracted high research interest for robot applications especially in oil and chemical industry and civil engineering. Robot body rolling while it travels within a pipe has been a problem for accurately collecting inspection data. Under certain circumstances where vision systems have to be employed, robot body rolling may cause vision inspection data to have little value as it is difficult to know where exactly the camera was looking at. This paper proposes an anti-rolling mechanism to hold consistent camera orientation. By changing the position angle of the robot legs, the mechanism is able to adjust the resistance to rolling within a pipe, therefore preventing robot rolling happening. The design makes use of the friction force caused by the gravity force of the robot to prevent the robot body rolling. The design analysis quantifies the effect of pipe radius, robot weight, payload, and payload offset distance in robot rolling. A test model was built based on the design concept. The experimental results obtained from the test model match the predication of the computational analysis. A real time vision system has been developed using FPGA and the algorithm of the structured laser light stripe configuration in the context of pipe inspection. The real-time hardware implementation of the algorithms on the robot itself removes the need to transmit raw video data back to an operator.