Abstract:The widespread availability of maps has enabled the development of numerous location-based applications, including navigation, ride-sharing, fitness tracking, gaming, robotics, and augmented reality. Today, the maps that power these services are predominantly controlled by a few large corporations and mostly cover outdoor spaces. As the use of these applications expands and indoor localization technologies advance, we are seeing the need for a scalable, federated location management system that can extend into private spaces. We introduce OpenFLAME (Open Federated Localization and Mapping Engine), the first federated and decentralized localization service. OpenFLAME links servers that handle localization for specific regions, providing applications with a seamless global view. Creating a federated localization system poses challenges, such as discovering the appropriate servers for a region and integrating services managed by independent providers. To address these issues and ensure scalability, we leverage Domain Name System (DNS) for service discovery and implement map abstractions to retrieve and merge locations across different maps. Our trace-driven study demonstrates that federated localization across remote servers is feasible with acceptable query latencies. To highlight the potential of the system, we developed an augmented reality navigation application for a large indoor space, showing that OpenFLAME can successfully power location-based applications.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the scale and privacy challenges encountered when existing centralized map services are extended to indoor spaces. Specifically: 1. **Limitations of Existing Map Services**: Current map services are mainly controlled by a few companies (such as Google and Apple) and mainly cover outdoor environments, relying on wide - area positioning systems such as GPS. However, with the development of indoor positioning technologies, such as machine - learning - based image matching and time - of - flight ranging (for example, compact LiDAR and UWB), these technologies are becoming standard features of smart phones, making accurate indoor positioning possible. 2. **Complexity and Sensitivity of Indoor Spaces**: The amount of work for storing and mapping indoor spaces is much greater than that for outdoor maps because the landmarks and features indoors are more dense. In addition, indoor maps contain sensitive information and need to be owned and controlled by the owners of the physical spaces. Many organizations (such as stores, companies, and government agencies) are unwilling to publicly host detailed indoor maps because this involves privacy issues. 3. **Deficiencies of Existing Commercial Solutions**: Existing commercial solutions (such as Google's Geospatial API, Niantic's Lightship VPS, and Apple's GeoTracking) although attempting to extend image - based positioning to a global level, have some significant drawbacks. These solutions need to pre - collect image data of the environment, construct 3D maps, and then perform device positioning according to the query images. These applications rely on specific service providers, cannot run seamlessly across the infrastructures of different service providers, and cannot promptly utilize the rapidly developing positioning technologies in academia. To solve these problems, the paper proposes OpenFLAME, which is the first federated positioning and mapping service. OpenFLAME aims to solve the above problems in the following ways: - **Federated and Decentralized Positioning Service**: OpenFLAME decomposes the positioning service into independent "map servers" which are deployed and maintained by different entities. In this way, each organization can independently manage its own map data without having to publicly host it to third - party services. - **Using DNS for Service Discovery**: OpenFLAME utilizes the Domain Name System (DNS) to discover map servers that provide positioning services. The rough location information of the device is converted into a series of DNS - compatible geo - domains, and the address list of map servers is obtained by querying DNS. This method does not need to rely on a central database, ensuring the scalability and privacy protection of the system. - **Map Abstraction and Consistency**: To deal with the inconsistency issues between different maps, OpenFLAME introduces a simple map abstraction - "waypoints". This abstraction allows applications to smoothly transition between different map boundaries without forcing all maps to be aligned to the same geographic coordinate system. In conclusion, the goal of OpenFLAME is to build a scalable, privacy - protected federated positioning and mapping service to support future location - aware applications, especially in indoor environments.

OpenFLAME: Building a large scale federated localization and mapping service

FLAME: Federated Learning Across Multi-device Environments

LSFB: A Low-cost and Scalable Framework for Building Large-Scale Localization Benchmark

Concurrent Flow-Based Localization and Mapping in Time-Invariant Flow Fields

Distributed Dynamic Map Fusion via Federated Learning for Intelligent Networked Vehicles

Federated Learning with Unsourced Random Access.

FLAME: Learning to Navigate with Multimodal LLM in Urban Environments

FLINT: A Platform for Federated Learning Integration

OpenFL: An open-source framework for Federated Learning

FL-AMM: Federated Learning Augmented Map Matching with Heterogeneous Cellular Moving Trajectories

FLAME: Adaptive and Reactive Concept Drift Mitigation for Federated Learning Deployments

StarFL: Hybrid Federated Learning Architecture for Smart Urban Computing

A Trustworthy AIoT-enabled Localization System via Federated Learning and Blockchain

Project Florida: Federated Learning Made Easy

Providing Location Information at Edge Networks: A Federated Learning-Based Approach

A Federated Learning Framework for Fingerprinting-Based Indoor Localization in Multibuilding and Multifloor Environments

SF-Loc: A Visual Mapping and Geo-Localization System based on Sparse Visual Structure Frames

Federated Learning-Based Localization with Heterogeneous Fingerprint Database

Seamless Fusion: Multi-Modal Localization for First Responders in Challenging Environments

Map Recovery and Fusion for Collaborative Augment Reality of Multiple Mobile Devices

Flexible LAN-WAN Orchestration for Communication Efficient Federated Learning over Large-Scale Mobile Devices