Underwater data caching with active and passive fiducial markers for asynchronous multi-ROV mapping and inspection in communication-challenged environments
Abstract
Approved for Public Release.Multi-robot teams require communication to coordinate their actions, cooperate, and share information. Yet, communicating between robots in the underwater domain remains challenging. Underwater acoustic modems achieve transmission rates of less than a kilobyte per second at long range (1000s of meters) and a few kilobytes per second at short range (less than 300 meters). This is a stark contrast to technologies such as Wireless Ethernet (WiFi) and 5G that provide transmission rates of megabytesper second when used on land and in the air. Indeed, communication underwater is 1000s of times slower than communication on land or in the air. We hypothesize that autonomous underwater vehicles (AUVs) can achieve new and better forms of coordination and cooperation in the underwater domain by storing data in the environment near to where that data is relevant. Storing information in the underwater environment can be achieved, for example, by leveraging recent breakthroughs in sonar fiducial technology and short-range light-based communication channels. We hypothesize that storing data in the underwater environment will enable new and better forms ofmulti-robot coordination and cooperation, especially for the problems of underwater multi-robot simultaneous localization and mapping (SLAM), and underwater multi-robot surveying and inspection.The proposed research effort will create novel methods (frameworks and algorithms) that use data storage in the environment to facilitate underwater multi-agent SLAM and underwater multi- agent surveying and inspection. The proposed effort will also create new theory (mathematical models and analysis tools) for understanding such methods. Placing passive artificial landmarksin an environment has already been used for underwater SLAM. We hypothesize that storingadditional data at mission relevant locations can further improve performance by reducing the needto for AUVs co-locate and/or synchronize their surfacing maneuvers for the purposes of data exchange. Storing data in situ will provide additional benefits in dynamic environments, e.g., formulti-robot SLAM and inspection missions that require understanding how the environment changes over time. Recent breakthroughs in underwater sonar fiducial marker technology make it possible to encode a few bytes of data into a passive fiducial marker that can be observed using an imaging sonar sensor. We will consider how using such passive forms of information storage can improve multi-robot algorithms by, e.g., storing mission-relevant data such as the geometric arrangementsof nearby landmarks.Other recent breakthroughs have demonstrated that light-based communication protocols can be used underwater to read/write gigabytes of information to waterproof digital data storage devices at short-range. We will consider how using such forms of active information storage can improve underwater multi-robot algorithms. For example, by storing gigabytes of algorithmic-relevant maps and historical navigation data in situ at mission relevant locations.In addition to the creation of novel multi-robot methods for SLAM and inspection that use underwater data storage for communication, the proposed research aims to develop a robust theoretical and empiricalunderstanding of these methods. We will quantify the costs and benefits of storing different amounts of data in the underwater environment. For example, by developing theory and running simulations to understand how mission performance is affected by the number of storage locations and the amount of data stored per location. Another goal of this effort is to understand and quantify how well such methods perform in situ. New methods will be tested and evaluated with Blue ROVs in existing underwater testbeds located at the University of Maryland,the US Naval Research Laboratory, and the US Naval Academy, and in underwater field tests near the US Naval Academy.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 12, 2025
- Source ID
- N000142512205
Entities
People
- Michael Otte
Organizations
- Office of Naval Research
- United States Navy
- University of Maryland