USVS IN SUPPORT OF MULTI-VEHICLE MULTI-DOMAIN MARITIME AUTONOMY
Abstract
It is proposed to build on the USV development effort at Florida Atlantic University in support of developing a system of collaborative multi-vehicle, multi-domain autonomous platforms for MCM, coastal surveillance, and protection of at-sea assets. The proposed effort will leverage/ collaborate with FAU’s ongoing NEEC effort in conjunction with NSWC, Panama City, that involves use of two UUVs for adaptive subsurface sensing in support of localization and classification of targets on the sea bottom. A significant challenge in such MCM operations is determining the exact geolocation of the vehicles as well as the observed targets. The proposed effort will aim to develop a collaborative system including a USV that will serve as an acoustic beacon for the two UUVs, thereby enabling the two UUVs to determine their geolocations relative to the USV, and hence their absolute geolocations though communication with the USV. Additionally, it is proposed to develop a collaborative multi-vehicle multi-domain system that may serve as an unmanned mobile network of nodes for coastal surveillance, or as a group of sentries around an at-sea asset. Under the current effort, a 16ft catamaran USV is being developed with capabilities for autonomous navigation, intelligent tracking of dynamic obstacles, and automated launch and recovery of an aerial drone and a UUV from the USV platform (Sinisterra et al., 2017b). The USV also serves as a docking station for power and data transfer between the USV and an aerial drone. The proposed effort will build on this work and the related software and hardware development, 1) to explore strategies for cooperation between a USV and UUVs in providing accurate geolocation of the UUVs in search for subsurface targets of interest; 2) to extend the current effort in software and hardware development to include capabilities for multi-sensor perception, collision avoidance, simultaneous localization and mapping (SLAM), and improved low-level control in adverse weather conditions. 3) to develop strategies for cooperative operations between teams of unmanned vehicles in accomplishing desired missions, using both modeling and simulation and at-sea tests; 4) to train and involve graduate and undergraduate students in unmanned autonomous marine vehicles-related research in support of NNRNE efforts. Implementation of the SLAM algorithm will provide a world map in support of situational awareness and improved navigation and path planning. A neural-network based low-level controller will be developed for maintaining planned trajectory under adverse environmental conditions. Further, high-resolution optical and infra-red (IR) devices will be implemented on the vehicle together with corresponding machine vision algorithms, including a tracking-learningdetection (TLD) algorithm, for detection, classification and tracking of dynamic objects of interest – detection could, for example, result in raising of an alert flag and trigger video streaming to a shore station for potential human intervention. Machine vision algorithm will also be implemented on the aerial drone, thereby broadening the synoptic and subsurface perception capability of the combined system, and enabling development of adaptive 3-D surveillance maps of the surroundings as well as improved detection and tracking of specific objects. The aim of the development approach is to foster open architecture and modularity to provide flexibility in the utility of the system in the maritime domain.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 26, 2018
- Source ID
- N000141812212
Entities
People
- Manhar Dhanak
Organizations
- Florida Atlantic University
- Office of Naval Research
- United States Navy