Abstract
Simultaneous localization and mapping (SLAM) in unmanned vehicles, such as drones, has great usability potential in versatile applications. When operating SLAM in multi-drone scenarios, collecting and sharing the map data and deriving converged maps are major issues (regarded as the bottleneck of the system). This paper presents a novel approach that utilizes the concepts of distributed ledger technology (DLT) for enabling the online map convergence of multiple drones without a centralized station. As DLT allows each agent to secure a collective database of valid transactions, DLT-powered SLAM can let each drone secure global 3D map data and utilize these data for navigation. However, block-based DLT—a so called blockchain—may not fit well to the multi-drone SLAM due to the restricted data structure, discrete consensus, and high power consumption. Thus, we designed a multi-drone SLAM system that constructs a DAG-based map database and sifts the noisy 3D points based on the DLT philosophy, named DAGmap. Considering the differences between currency transactions and data constructions, we designed a new strategy for data organization, validation, and a consensus framework under the philosophy of DAG-based DLT. We carried out a numerical analysis of the proposed system with an off-the-shelf camera and drones.
Original language | English |
---|---|
Article number | 34 |
Journal | Drones |
Volume | 6 |
Issue number | 2 |
DOIs | |
Publication status | Published - 2022 Feb |
Bibliographical note
Publisher Copyright:© 2022 by the authors. Licensee MDPI, Basel, Switzerland.
Keywords
- DAG-based DLT
- Distributed ledger technology
- Multi-agent SLAM
ASJC Scopus subject areas
- Control and Systems Engineering
- Information Systems
- Aerospace Engineering
- Computer Science Applications
- Artificial Intelligence