---
title: 'AUV-Fusion: Underwater Sensor Fusion'
url: https://www.emergentmind.com/topics/auv-fusion
type: topic
---

# AUV-Fusion: Underwater Sensor Fusion

Searching arXiv for recent papers using the term “AUV-Fusion” and closely related underwater sensor-fusion frameworks.
arxiv_search(query="AUV-Fusion autonomous underwater vehicle sensor fusion", max_results=10, sort_by="relevance")
AUV-Fusion is a label used in recent arXiv literature for several, partly overlapping lines of work on autonomous underwater vehicle sensor fusion. In its narrow sense, it denotes an INS/DVL navigation framework that augments a nonlinear error-state EKF with DVL-derived acceleration updates to improve convergence and bias estimation [2308.11762]. In a broader sense, it names a research program spanning classical model-based filtering, AI-adaptive navigation, tightly coupled visual–acoustic–inertial odometry, and opti-acoustic volumetric mapping for AUVs operating without GNSS and under visibility, synchronization, and sensor-availability constraints [2605.04672, 2603.14457]. The same name is also used in an unrelated cross-modal adversarial attack framework for visual-aware recommender systems, which is a distinct topic and not part of underwater robotics [2507.22880].

## 1. Terminological scope and research setting

Within underwater robotics, AUV-Fusion refers to the integration of heterogeneous sensing modalities—most often INS, DVL, cameras, sonar, pressure sensors, magnetic compasses, LBL, GNSS surface fixes, and IMU-derived preintegration—into a single estimation or mapping pipeline. The stated motivation is consistent across the literature: electromagnetic signals are unavailable underwater, DVL measurements may become incomplete or unavailable, visual sensing is highly sensitive to turbidity and low contrast, and asynchronous sensors complicate tightly coupled estimation [2605.04672, 2404.13742, 2603.14457].

The term has both framework-specific and umbrella usages. In “INS/DVL Fusion with DVL Based Acceleration Measurements,” the integrated navigation method itself is described as “AUV-Fusion” and is centered on software-only modification of an INS/DVL EKF [2308.11762]. In “AI-Aided Advancements in Autonomous Underwater Vehicle Navigation,” “AUV-Fusion” is used as a compendium-level designation for a taxonomy of classical filters, learning-based calibration, adaptive filtering, visual–inertial fusion, and emerging reinforcement-learning-driven sensor weighting [2605.04672]. This suggests that the term functions less as a single standardized algorithmic object than as a family resemblance across underwater fusion architectures.

A concise cross-section of the literature is given below.

| Paper | Modalities | Principal focus |
|---|---|---|
| [2308.11762] | INS + DVL | DVL-based acceleration update inside EKF |
| [2404.13742] | INS + DVL | Missing-beam regression with HNC for LC/TC fusion |
| [2303.17005] | Vision + DVL + IMU + Pressure | Tightly coupled MSCKF odometry |
| [2310.14163] | SINS + LBL + DVL + MCP + PS + GNSS | Factor-graph integrated navigation |
| [2603.14457] | Stereo sonar + monocular camera | Confidence-weighted GP volumetric mapping |
| [2512.05303] | LiDAR + IMU + dual orthogonal FLS | Seabed-to-sky factor-graph mapping |

## 2. INS/DVL-centered navigation formulations

A core AUV-Fusion lineage is the INS/DVL error-state filter. The canonical state comprises position, velocity, attitude or misalignment, and inertial biases, while DVL velocity enters as an aiding measurement. In the acceleration-augmented variant, the error-state is
\[
\delta x =
\bigl[\;\delta v^n\,,\,\phi^n\,,\,b_a\,,\,b_g\;\bigr]^T
\in\mathbb R^{12},
\]
with continuous-time linearized dynamics
\[
\dot{\delta x}
= F\,\delta x + G\,w.
\]
The distinctive step is the introduction of a DVL-derived acceleration measurement obtained from a least-squares fit over recent DVL velocity samples. With
\[
V = (L^T L)^{-1} L^T\,V_h,
\]
the acceleration estimate is extracted from the second row, and the EKF is updated using
\[
\delta z_a
= \tilde a^b \;-\; R^b_d\,\hat a^d,
\qquad
H_a
= \bigl[\,0_{3\times3}\;\;-\;R^b_n\,[g^n\times]\;\;I_3\;\;0_{3\times3}\bigr].
\]
The reported effect is not the creation of new observable modes—the paper explicitly states that the unobservable null-space is unchanged—but faster convergence of accelerometer and gyroscope residual errors and lower residual bias error [2308.11762].

This INS/DVL-centered thread has diversified in several directions. A hybrid adaptive velocity-aided navigation filter learns the momentary process-noise covariance \(Q_k\) from handcrafted IMU features and inserts the learned covariance into an es-EKF, reporting SRMSE \(0.980\) m/s and SMAE \(0.866\) m/s for the hybrid HCF-Ensemble, compared with \(1.082\) m/s and \(0.961\) m/s for a constant nominal \(Q\) baseline [2211.01329]. A DVL outage formulation, ST-BeamsNet, regresses body-frame velocity during complete DVL outage from inertial windows and previous DVL velocity history using a Set-Transformer, reporting RMSE \(0.098\) m/s, MAE \(0.064\) m/s, \(R^2=0.978\), VAF \(97.997\), and an \(8.574\%\) speed RMSE, which is approximately \(26\%\) better than the moving-average approach [2212.11671]. A complementary missing-beam framework, HNC, regresses two or three absent Janus beams and reinserts them into loosely or tightly coupled EKF updates; for two missing beams it reports VRMSE \(0.17\) m/s on both trajectories for HNLC and HNTC, and for three missing beams \(0.25\) m/s and \(0.29\) m/s, with average performance gains reported as \(96.15\%\) relative to baseline model-based approaches and \(12.41\%\) relative to a model-based beam estimator [2404.13742].

A separate but adjacent issue is INS/DVL alignment. AlignNet formalizes alignment as direct regression from synchronized \([v^b;v^d]\) windows to Euler-angle offsets \([\hat\phi,\hat\theta,\hat\psi]\). On synthetic lawn-mower trajectories, it reports RMSE \(5.24^\circ\) at \(T=5\) s, \(4.72^\circ\) at \(25\) s, \(3.64^\circ\) at \(50\) s, and \(2.93^\circ\) at \(100\) s, with convergence to approximately \(3.5^\circ\) RMSE in approximately \(25\) s versus approximately \(100\) s for the velocity-based SVD baseline [2503.21350]. This suggests that, in the AUV-Fusion literature, calibration and estimator adaptation are treated as integral parts of fusion rather than merely preprocessing.

## 3. Tightly coupled odometry and smoothing back-ends

A second major branch of AUV-Fusion emphasizes tightly coupled multi-sensor odometry. In the under-ice visual–DVL–IMU–pressure framework, visual features, DVL velocity, IMU mechanization, and pressure depth are integrated within a Multi-State Constraint Kalman Filter. The state is
\[
\mathbf{x}_k =
\begin{bmatrix}
\mathbf{x}_{\mathrm{IMU}}\\
\mathbf{x}_{\mathrm{Clone}}
\end{bmatrix},
\quad
\mathbf{x}_{\mathrm{IMU}}\in\mathbb{R}^{15},\;
\mathbf{x}_{\mathrm{Clone}}\in\mathbb{R}^{6n},
\]
with adaptive keyframe cloning and a DVL-aided feature enhancement that uses sparse DVL point clouds to correct visual feature depth under short baselines. In a frozen Keweenaw Waterway dataset of about \(200\) m, plain MSCKF plus vision, DVL, and pressure achieved RMSE\(_{XY}\) approximately \(1.71\) m, adding keyframe clones reduced it to approximately \(1.14\) m, and the full system produced best ATE\(_{XY}\) approximately \(1.11\) m, while monocular VIO alone failed under prolonged hovering and DVL-IMU-pressure dead reckoning drifted unbounded [2303.17005].

Factor-graph smoothing provides a distinct fusion regime. FGO-ILNS represents each state as
\[
x_i = [p_i, v_i, q_i, b_{a,i}, b_{\omega,i}],
\]
and constructs a sliding-window graph containing IMU preintegration factors, floating-LBL slant-range difference factors, GNSS, DVL, MCP, PS, and a marginalization prior. Asynchronous sensors are handled by forward-backward IMU preintegration, and history outside the window is compressed by Schur complement. The optimization solves
\[
X^* = \arg \min_X \sum_i\|r_i(X)\|_{\Lambda_i}^2
\]
using Gauss–Newton or Levenberg–Marquardt in Ceres Solver. In simulation, the best ENU RMS at a \(20\) s sliding window was East \(0.67\) m, North \(0.28\) m, Up \(0.01\) m; under aggressive maneuvers, EKF drift exceeded \(16\) m while FGO stayed below \(1\) m; and in an outage scenario the maximum drift remained below \(5\) m for FGO versus above \(50\) m for a federated EKF [2310.14163].

FAR-AVIO advances the tightly coupled formulation by embedding a Schur-complement landmark elimination inside an EKF. Its nominal state augments pose, velocity, inertial biases, and both camera and DVL extrinsics,
\[
x_b =
\begin{bmatrix}
p_b^w\ v_b^w\ R_b^w\ b_a\ b_g\ T_c^b\ T_D^b
\end{bmatrix},
\]
and the minimal error state has \(n_x=27\). Visual updates are reduced to fixed-size equivalent observations by eliminating landmark states through
\[
\bar H = C_1 - C_2\,C_3^{-1}C_2^\top,
\]
which yields constant-time EKF updates. The framework adds AWARE, an online reliability mechanism that scales measurement covariances and can disable or re-enable a sensor stream based on quality queues and thresholds, and it performs online DVL–IMU extrinsic calibration by including \(T_D^b\) in the state. On eight tank sequences, FAR-AVIO reports overall RMSE approximately \(2.08\) m versus \(2.42\) m for AQUA-SLAM and more than \(6\) m for ORB-SLAM3 and VINS-Fusion; AWARE improves RMSE by \(10\)–\(25\%\); online extrinsic calibration converges to below \(0.05\) rad and below \(0.05\) m; and runtime on Jetson Orin NX is \(28\) ms/frame versus \(62\) ms for VINS-Fusion [2512.20355].

## 4. Opti-acoustic fusion and volumetric mapping

Another AUV-Fusion trajectory moves from state estimation to geometry reconstruction. “Towards Versatile Opti-Acoustic Sensor Fusion and Volumetric Mapping” proposes a BlueROV2 platform equipped with a forward-looking monocular camera and two orthogonally mounted multibeam imaging sonars: a “horizontal” sonar with \(130^\circ\) H \(\times\) \(20^\circ\) V and a “vertical” sonar with \(20^\circ\) H \(\times\) \(130^\circ\) V. Overlapping sonar fan-planes resolve elevation ambiguity in the intersection region, while the camera provides elevation cues through ROI segmentation using YOLO11n-seg. Stereo-sonar points, sonar-to-image projection points, and image-expansion points are each assigned confidence weights \(\alpha_{ss}\), \(\alpha_s\), and \(\alpha_e\), with heteroscedastic GP noise
\[
\sigma_{n,i}^2=\frac1{\alpha_i}.
\]
Occupancy is then estimated by Gaussian Process Volumetric Mapping with a Matérn \(\nu=3/2\) kernel,
\[
k(d)=\sigma_f^2\Bigl(1+\tfrac{\sqrt3d}{l}\Bigr)e^{-\sqrt3d/l},
\]
and converted to occupancy probability via
\[
p(\mathrm{occ}\mid\mathbf{x}_j) =\frac{1}{1+\exp\bigl(-\gamma\,\omega_j\bigr)}.
\]
The reported voxel resolution is \(2.5\) cm, keyframes are triggered every \(5\) cm or \(10^\circ\), confidence-weighted GP mapping adds approximately \(0.055\) s/frame relative to standard GP, and OctoMap is more than \(2\times\) slower. In tank experiments, the proposed GPC SS RGB achieved approximately \(1.17\) cm MAE for a single disk and \(1.62\) cm for a double disk, together with the lowest RMSE and precision above \(80\%\); in a turbid marina, it qualitatively captured wooden pilings and small front-pipe features, while GP-only and OctoMap baselines either lacked coverage or overestimated occupancy [2603.14457].

A related but surface-oriented extension fuses a downward-looking 3D LiDAR and IMU with a dual orthogonal forward-looking sonar pair in a modified LIO-SAM back-end. Stereo-derived 3D sonar points and leading-edge line scans are inserted into a single factor graph through motion-interpolated poses between LiDAR keyframes. The system reports approximately \(2.65\) Hz map updates and approximately \(2.85\) Hz odometry, above-water Euclidean errors of \(0.139\)–\(0.198\) m after rigid alignment to UTM32, and underwater wall-normal cosine similarity of approximately \(0.991\)–\(0.993\), while noting a small systematic underestimation and lateral offset due to residual extrinsic errors and elevation uncertainty outside the stereo overlap [2512.05303]. Although this platform is an autonomous surface vehicle rather than an AUV, it extends the same orthogonal-sonar and factor-graph principles into unified maritime mapping.

## 5. Mathematical patterns across the literature

Across these works, AUV-Fusion is less a single estimator than a recurring set of mathematical design patterns. The most basic layer is model-based Bayesian filtering. The compendium literature distinguishes EKF, UKF, and PF as the canonical classical architectures. For EKF, the prediction and measurement update are written as
\[
\hat x_{k|k-1} = f(\hat x_{k-1|k-1}, u_{k-1}), \qquad
P_{k|k-1} = F_{k-1} P_{k-1|k-1} F_{k-1}^\top + Q_{k-1},
\]
\[
K_k = P_{k|k-1} H_k^\top (H_k P_{k|k-1} H_k^\top + R_k)^{-1},
\qquad
\hat x_{k|k} = \hat x_{k|k-1} + K_k (z_k - h(\hat x_{k|k-1})),
\]
and adaptive weighting may update \(R_k\) by covariance matching [2605.04672]. This formulation underlies the INS/DVL acceleration update, missing-beam HNC, and several visual–DVL–inertial systems.

A second pattern is graph-based smoothing with preintegration and marginalization. FGO-ILNS uses forward-backward preintegration to project asynchronous sensors onto adjacent IMU states and applies a Schur-complement prior when sliding the window [2310.14163]. FAR-AVIO instead moves Schur elimination into the EKF update itself, achieving fixed state size and constant-time updates [2512.20355]. This suggests that the boundary between “filter” and “optimizer” has become porous in modern underwater fusion.

A third pattern is learned adaptation or learned measurement completion. Examples include bagged-tree estimation of process-noise covariance from IMU statistics [2211.01329], Set-Transformer regression of velocity during complete DVL outage [2212.11671], HNC beam completion for limited DVL measurements [2404.13742], and 1D-CNN alignment regression for \(R_b^d\) estimation [2503.21350]. The AI-aided survey additionally places these alongside ResAlignNet, A-KIT, ProcessNet, Gaussian-process covariance adaptation, PiDR, and prospective actor–critic tuning of filter covariances or sensor weights [2605.04672]. A plausible implication is that AUV-Fusion research increasingly treats uncertainty tuning, alignment, and sensor availability as learning problems embedded around classical estimators rather than replacements for them.

## 6. Empirical behavior, limitations, and boundaries of the term

The reported performance gains in AUV-Fusion papers are substantial but method-specific. The acceleration-augmented INS/DVL EKF reports Z-axis accelerometer bias error reductions of \(53\%\), X and Y gyro bias error reductions of \(48\%\), leveling-angle reductions of \(7\%\), average improvement of approximately \(19\%\) in a straight run, and average improvement of approximately \(40\%\) in a figure-eight experiment, with convergence-time reductions of \(57\%\) and \(54\%\), respectively [2308.11762]. HNC reports seamless operation under two- or three-beam loss [2404.13742]. FGO-ILNS reports greater than \(97\%\) horizontal-accuracy improvement under complex dynamics relative to EKF and maintains less than \(5\) m maximum drift in outage scenarios [2310.14163]. Opti-acoustic GP mapping reports lowest MAE and RMSE among the compared baselines while maintaining substantial coverage in both clear and turbid water [2603.14457].

At the same time, the limitations are explicit. The DVL acceleration update “does not introduce new observable modes” and “does not rectify classical unobservable biases (e.g. heading drift)” [2308.11762]. AlignNet is evaluated on high-fidelity simulation and may face degradation from multipath acoustic effects and irregular seafloor in real deployment [2503.21350]. The AI-aided overview emphasizes time synchronization, visibility and feature degradation in turbid water, computational constraints on onboard CPUs, generalization across vehicles and sensor grades, and the need to balance real-time accuracy with energy and latency [2605.04672]. The dual-sonar/LiDAR mapping paper attributes systematic underwater offsets to residual extrinsic errors and elevation uncertainty outside stereo overlap [2512.05303]. These are not peripheral issues; they are recurrent structural constraints on underwater fusion.

Finally, the term’s boundaries matter. “AUV-Fusion: Cross-Modal Adversarial Fusion of User Interactions and Visual Perturbations Against VARS” is a recommender-systems attack framework that models user preference embeddings and injects perturbations into a diffusion-VAE latent space; despite the shared name, it does not concern autonomous underwater vehicles, inertial navigation, or marine sensing [2507.22880]. For the underwater literature, AUV-Fusion therefore denotes a technical tradition centered on multi-sensor integration for navigation, odometry, and mapping, rather than a single universally defined algorithm.

Source: https://www.emergentmind.com/topics/auv-fusion