Positioning and control of an AUV inside a water pipeline for non-contact in-service inspection

Water pipelines are an important part of our social infrastructure. Early failure detection of the pipelines will help to prevent breakdown and avoid the large cost involved if a accident occurs. As pipelines are enclosed environments and difficult to access, pipeline inspection robots are increasingly used for routine inspection and early failure detection. In our research we propose the use of a small untethered AUV for the visual inspection of the pipeline interior. The vehicle carries optical sensors to aid non-contact motion inside the pipeline and also to capture images of the pipeline interior. The vehicle has identical front and rear sections and has a pair of cone laser and fish eye camera on either ends. The cone lasers and cameras act as range sensors as well as payload to capture pipe line interior images. A realtime vision based control system was developed for the autonomous operation of the proposed vehicle inside pipelines. A new offline position estimation algorithm estimates lateral position and pose of the vehicle inside pipeline. Laboratory experiments were performed in a test pipeline without and with water flow. The results of the experiment show that the proposed technique is efficient for non-contact inspection of a pipeline and helps to make vehicle compact to navigate through the enclosed environment inside pipelines.