Low-Cost Underwater In-Pipe Centering and Inspection Using a Minimal-Sensing Robot

TL;DR

This work presents a minimal-sensing framework for autonomous underwater in-pipe navigation using only an IMU, a pressure sensor, and two sonar measurements to center and traverse a submerged pipe of known radius. It combines a computationally light sonar-range estimation pipeline, a closed-form two-point geometry for pipe-centre estimation, and an uncertainty-aware, PD-based control strategy that scales commands by measurement confidence. Validation spans real-sensor wall-range accuracy (RMSE ≈ 0.02 m), 2D simulations showing rapid convergence to the pipe centre, and field experiments with a deformable fabric culvert demonstrating stable centering and complete traversal. The approach highlights that low-cost sensing and processing can enable practical autonomous underwater inspection in confined geometries, reducing hardware and computation requirements while preserving robustness.

Abstract

Autonomous underwater inspection of submerged pipelines is challenging due to confined geometries, turbidity, and the scarcity of reliable localization cues. This paper presents a minimal-sensing strategy that enables a free-swimming underwater robot to center itself and traverse a flooded pipe of known radius using only an IMU, a pressure sensor, and two sonars: a downward-facing single-beam sonar and a rotating 360 degree sonar. We introduce a computationally efficient method for extracting range estimates from single-beam sonar intensity data, enabling reliable wall detection in noisy and reverberant conditions. A closed-form geometric model leverages the two sonar ranges to estimate the pipe center, and an adaptive, confidence-weighted proportional-derivative (PD) controller maintains alignment during traversal. The system requires no Doppler velocity log, external tracking, or complex multi-sensor arrays. Experiments in a submerged 46 cm-diameter pipe using a Blue Robotics BlueROV2 heavy remotely operated vehicle demonstrate stable centering and successful full-pipe traversal despite ambient flow and structural deformations. These results show that reliable in-pipe navigation and inspection can be achieved with a lightweight, computationally efficient sensing and processing architecture, advancing the practicality of autonomous underwater inspection in confined environments.

Figures (7)

• Figure 1: The Blue Robotics BlueROV2 Heavy vehicle used for experiments, equipped with a Ping360 rotating sonar mounted on the nose for 360° scanning and a Ping1D downward-facing sonar for bottom range measurement.
• Figure 2: Sonar processing pipeline for range estimation. Sonar intensity profiles are processed through near-field suppression, denoising, and edge enhancement. The first significant return is detected as the pipe wall echo, converted to a range value, and smoothed over time to produce the final distance estimates for real-time navigation.
• Figure 3: Geometric construction of candidate centers. Two sonar returns, $\mathbf{p}_{\mathrm{D}}$ and $\mathbf{p}_{360}$, define a chord on the circular pipe wall. The chord midpoint is computed, and two candidate circle centers ($\mathbf{c}_1$, $\mathbf{c}_2$) are obtained along the perpendicular bisector. The algorithm selects the physically consistent center based on geometric and temporal criteria.
• Figure 5: Example adaptive measurement covariance as a function of beam-separation angle $\theta$. Plot of the modeled measurement variance $\sigma^2(\theta)$ showing reduced uncertainty near orthogonal alignment and increased uncertainty when beams are nearly colinear. The covariance function governs adaptive weighting in the Kalman filter.
• Figure 6: Comparison between the ground-truth wall distances (solid blue) and the estimated wall-range measurements (dashed orange) across sonar intensity profiles. The method reliably identifies the first wall return, producing range estimates that closely match ground truth, with a root mean square error (RMSE) of 0.02 m.