24 May 2022 in IEEE Journal of Oceanic Engineering
Many modern imaging sensors must obtain multiple looks or “views” of a target at different orientations to automatically classify it with high confidence. Therefore, when tasked with classifying many targets, a mobile sensor may need to travel a long distance to change its position and orientation relative to every target, resulting in costly and time-consuming operations. This article presents a novel and general approach, referred to as informative multiview planning (IMVP) that simultaneously determines the most informative sequence of views and the shortest path between them. The approach is demonstrated on an underwater multitarget classification problem in which a sidescan sonar installed on an unmanned underwater vehicle must classify all targets in minimum time. Simulation and experimental results show that IMVP can achieve the same, or better, classification performance in half the time of existing multiview path planning methods. Also, by determining the most informative views and the shortest path between them, IMVP significantly improves classification efficiency, classification confidence level, as well as performance robustness.
3 March 2025 in IEEE Robotics and Automation Letters
In this letter, we propose a point cloud structural similarity-based loop detection method for underwater Simultaneous Localization and Mapping using sonar sensors. Existing sonar-based loop detection approaches often rely on 2D projection and keypoint extraction, which can lead to data loss and poor performance in feature-scarce environments. Additionally, methods based on neural networks or Bag-of-Words require extensive preprocessing, such as model training or vocabulary creation, reducing adaptability to new environments. To address these challenges, our method directly utilizes 3D sonar point clouds without projection and computes point-wise structural feature maps based on geometry, normals, and curvature. By leveraging rotation-invariant similarity comparisons, the proposed approach eliminates the need for keypoint detection and ensures robust loop detection across diverse underwater terrains. We validate our method using two real-world datasets: the Antarctica dataset obtained from deep underwater and the Seaward dataset collected from rivers and lakes. Experimental results show that our method achieves the highest loop detection performance compared to existing keypoint-based and learning-based approaches while requiring no additional training or preprocessing.
25 April 2024 in Journal of Intelligent & Robotic Systems
Search operations performed by adaptive autonomous maritime vehicles have been a topic of considerable interest for many years. Such operations require carefully scheduled coordination of multiple vehicles performing search tasks across the region of interest. Due to the inherent uncertainty of the maritime environment, however, an initially planned search schedule may not be maintained if the vehicles have significant capability to adapt their tasks to match the environment they detect in real time. We propose a multi-vehicle adaptive algorithm for dynamic evaluation and elastic re-planning of variable-length tasks commonly found in the maritime environments. In adaptive evaluation and re-planning problems, a set of tasks are initially planned for execution by adaptive, autonomous search vehicles. Tasks are allocated to search vehicles under a pre-defined schedule based on prior knowledge and desired outcome. Because of the vehicles’ autonomy and reactivity to in situ conditions such as environment or target pose, the precise duration and actions required by each task are unknown a priori. We develop a hidden Markov model (HMM) for propagating task estimates, coupled with a quadratic-programming-based elastic re-scheduler. The result is an integrated estimate-and-schedule adaptation scheme that quickly and efficiently re-plans the vehicles’ schedules based on in situ observations. The numerical simulation results show that this novel HMM approach decreases avoidable schedule variation by over a factor of two compared to existing methods.
15 August 2024 in IEEE Access
This paper presents a formulation for deterministically calculating optimized paths for a multi-agent system consisting of heterogeneous vehicles. The key idea is the calculation of the shortest time for each agent to reach every grid point from its known initial position. Such arrival time map is efficiently computed using the Fast Marching Method (FMM), a computational algorithm originally designed for solving boundary value problems of the Eikonal equation. By leveraging the FMM, we demonstrate that the minimal time rendezvous point and paths for all member vehicles can be uniquely determined with minimal computational overhead. The scalability and adaptability of the present method during online execution are investigated, followed by a comparison with a baseline method that highlights the effectiveness of the proposed approach. Then, the potential of the present method is showcased through a virtual rendezvous scenario involving the coordination of a ship, an underwater vehicle, an aerial vehicle, and a ground vehicle, all converging at the optimal location within the Tampa Bay area in minimal time. The results show that the developed framework can efficiently construct continuous paths of heterogeneous vehicles by accommodating operational constraints via an FMM algorithm.
11 February 2026 in IEEE Access
Localization in Global Navigation Satellite System-denied environments is critical for autonomous systems, where traditional simultaneous localization and mapping methods often lack generalizability. This paper presents an uncertainty-aware magnetic navigation framework that directly approximates information from the localization distribution to optimize trajectory selection. By minimizing information loss while ensuring goal attainment, it achieves improved accuracy and efficiency. Real-time simulation in a Global Navigation Satellite System-denied indoor environment, along with hardware validation, shows a localization performance of 5% higher, 34% lower failure rates, and more than 99% computational savings compared to state-of-the-art methods, with consistent results across different anomaly maps. These findings highlight the practicality and robustness of the solution for long-duration magnetic-based navigation.