1 of 1

Active Sonar-Driven Iceberg Surface Following Path Planner for �Autonomous Underwater Vehicles�Tony Jacob

Department of Ocean Engineering, University of Rhode Island

Abstract:

Underwater portion of icebergs (about 88-90% of the overall shape) has to be measured to better quantify its climate impacts [1]. Advanced autonomy is needed for a safe AUV operation and produce high quality mapping products including iceberg rendering & surrounding water properties [2]. To this end, in this paper, we present an novel approach for generating guidance path for an AUV, based on its Occupancy Grid Map (OGM) constructed from the measurements from a Mechanical Scanning Imaging Sonar (MSIS). The method consists of three components: MSIS data pre-processing and costmap processing, generation of continuous path and heuristic based waypoint selection, to realize the autonomous iceberg wall-following behaviour.�

Keywords—Autonomous Underwater Vehicles (AUVs), Path Planning, Iceberg Mapping, Mechanical Scanning Imaging Sonar (MSIS), Occupancy Grid Maps (OGM)

1. MSIS Preprocessing�

Converting MSIS returns to Point Clouds.

�

Filtering the points: �1) Range threshold to remove ringing effects. 2) Extracting high intensity points using mean & standard deviation.

2. Costmap Processing

The filtered point-clouds are projected into an Occupancy Grid Maps to represent the surroundings.
Applying a series of image processing techniques to extract iceberg contour.

4. Path Following

Following Mode:

�

Validation & Conclusion:

The online path planner has been validated in Stonefish [3] on an iceberg drifting with a northward speed of 0.05 m/s, an eastward speed of 0.02 m/s and a rotational speed of 0.025 deg/s.

�-The vehicle was able to successfully circumnavigate at a standoff distance of 20m and tracking RMSE of 3m.

Figure 5. Overview of the mission where AUV (path in blue line) has circumnavigated around a moving iceberg once. Results from two other icebergs can be found in the paper.

Acknowledgements: This work is supported by NSF award #2221676 and the Smart Ocean Systems Lab, University of Rhode Island. We thank ONR for the grant to participate in this conference.

Figure 2. (A) The raw image from the OGM, (B) The resulting image after dilation, (C) After applying Canny Edge algorithm, (D) The final estimation of the iceberg surface contour in the processed image.

Figure 6. Standoff distance between the iceberg and the vehicle over time compared to the desired standoff value. The gray regions denote the times where the vehicle switched to Iceberg Reacquisition Mode.

Figure 4. Left : Simulated world (Stonefish). Right: Generated path from different points of the iceberg, visualized in Rviz.

Figure 3. The series of transformations between coordinate frames to generate the path.

Figure 1. Left: Point Cloud representation of the MSIS data. Each ping produces a distribution of intensity over range. Zoom-in picture is the returns due to ringing effects. Right: Filtered pointcloud

3. Path Generation

Define a local iceberg contour coordinate frame based on the trend of the contour.
Offset contour towards the vehicle with the desired stand-off distance.
Apply a curve fit to parameterize the offsetted points into a polynomial equation.
Generate waypoints from the polynomial.

Iceberg Reacquistion Mode:

For all points in {L}, distance to the point from the vehicle (d^L_V,i), the vehicle to point angle (ψ^L_V,i), the track angle (ψ^L_i,i+1), maximum surge velocity (u_max) and maximum yaw rate (r_max) of the vehicle are used to determine the cost of each point.��
The point with the minimum cost that falls between (-π/2, π/2) of the line frame {L} is then chosen and fed to the guidance system.�
This heuristic enables the vehicle to track the path with minimum time to intercept during clockwise iceberg circumnavigation.�
When the iceberg disappears from the FOV of the sonar (at sharp corners or concave features), Iceberg Reacquisition Mode is designed to gradually turn the vehicle 90 degrees clockwise to regain acoustic contacts.

Demo

Co-Authors: Mingxi Zhou, �Graduate School of Oceanography, University of Rhode Island

References: [1]: Y. Fang, “The impact of iceberg melting on the climate of the arctic circle,” in Proceedings of the 2022 6th International Seminar on Education, Management and Social Sciences (ISEMSS 2022).

[2]: P. W. Kimball and S. M. Rock, “Mapping of translating, rotating icebergs with an autonomous underwater vehicle,” IEEE Journal of Oceanic Engineering, vol. 40, no. 1, pp. 196–208, 2015.

[3]: P. Cie´slak, “Stonefish: An advanced open-source simulation tool designed for marine robotics, with a ros interface,” in OCEANS 2019 - Marseille, 2019, pp. 1–6.