Robust Underwater Localization of Buoyancy Driven microFloats Using Acoustic Time-of-Flight Measurements

Accurate underwater localization remains a challenge for inexpensive autonomous platforms that require highfrequency position updates. In this paper, we present a robust, low-cost localization pipeline for buoyancy-driven microFloats operating in coastal waters. We build upon previous work by introducing a bidirectional acoustic Time-of-Flight (ToF) localization framework, which incorporates both float-to-buoy and buoy-to-float transmissions, thereby increasing the number of usable measurements. The method integrates nonlinear trilateration with a filtering of computed position estimates based on geometric cost and Cramer-Rao Lower Bounds (CRLB). This approach removes outliers caused by multipath effects and other acoustic errors from the ToF estimation and improves localization robustness without relying on heavy smoothing. We validate the framework in two field deployments in Puget Sound, Washington, USA. The localization pipeline achieves median positioning errors below 4 m relative to GPS positions. The filtering technique shows a reduction in mean error from 139.29 m to 12.07 m, and improved alignment of trajectories with GPS paths. Additionally, we demonstrate a Time-Difference-of-Arrival (TDoA) localization for unrecovered floats that were transmitting during the experiment. Range-based acoustic localization techniques are widely used and generally agnostic to hardware-this work aims to maximize their utility by improving positioning frequency and robustness through careful algorithmic design.

💡 Research Summary

This paper presents a robust and comprehensive localization pipeline designed for inexpensive, buoyancy-driven microFloats (µFloats) operating in challenging coastal waters. The core challenge addressed is providing high-frequency, accurate position updates for these Lagrangian drifters without access to GPS underwater, using only low-cost acoustic nanomodems (Succorfish Delphis v3).

The authors’ primary innovation is the introduction of a bidirectional acoustic Time-of-Flight (ToF) framework. Moving beyond prior unidirectional (surface buoy-to-float) methods, the new system utilizes both uplink (float-transmit, buoy-receive) and downlink (buoy-transmit, float-receive) acoustic pings scheduled via TDMA. This doubles the potential measurement opportunities within a given time window, increasing data density and providing redundancy against acoustic path blockages, thereby enhancing overall system robustness.

The localization pipeline processes post-deployment data through a sequence of steps: acoustic ping matching and range calculation, depth compensation using the float’s pressure sensor to derive horizontal distances, initial outlier removal based on physical plausibility (e.g., velocity checks), and grouping of pings within a 5-second sliding window. Groups containing pings from at least three unique Surface Localization Buoys (SLBs) are then processed using nonlinear least-squares trilateration to estimate a 2D position.

A key contribution is the novel filtering and validation stage designed to reject erroneous position estimates without resorting to heavy temporal smoothing, which can obscure genuine high-frequency motion. This stage employs a dual-criterion filter: 1) it checks the residual cost from the trilateration optimization, and 2) it calculates and thresholds the theoretical positional uncertainty derived from the Cramér-Rao Lower Bound (CRLB). The CRLB, computed via the Fisher Information Matrix (FIM), provides a geometry-aware lower bound on estimation variance based on the configuration of the buoys and assumed measurement noise. Estimates with high residual costs or excessively high CRLB-derived uncertainties (σ_x, σ_y) are discarded as outliers. This method selectively removes errors caused by multipath, noise, or poor geometric dilution of precision (GDOP), while preserving accurate high-rate position fixes.

The framework was validated with data from two field deployments in Puget Sound, Washington. The results demonstrated a median positioning error of less than 4 meters relative to GPS-derived ground truth when the floats surfaced. The effectiveness of the CRLB/cost filtering was stark: it reduced the mean localization error from 139.29 m to 12.07 m, a ~91% improvement, and produced trajectories that aligned much more closely with the actual GPS paths. Furthermore, the authors demonstrated a fallback Time-Difference-of-Arrival (TDoA) localization method for floats that were actively transmitting but not recovered, using only the arrival times of their signals at multiple SLBs, processed through the same filtering logic.

In summary, this work successfully demonstrates that careful algorithmic design—specifically, bidirectional measurement fusion and theoretically grounded, geometry-aware outlier rejection—can dramatically improve the accuracy and robustness of underwater localization for low-cost platforms, making dense, accurate swarm tracking in dynamic coastal environments a practical reality.