BLE beacon indoor localization dataset

This dataset was created to facilitate research into indoor localization with BLE beacons. Data was collected from September 2018 to May 2019 in two separate locations. Several participants assisted with the experiment each carrying a BLE beacon and a smartphone. This dataset is released in hope that researchers will use it as a benchmark for indoor localization techniques. We initially collected this dataset noting a lack of available indoor localization datasets with labeled data. Our implementation uses the fixed-receiver moving-transmitter method of indoor localization. We use Raspberry Pis and BLE beacons as our receivers and transmitters respectively. Android Minimum Version | 4.4 Beacon Type | Gimbal Series 10 Beacon Profile | iBeacon Beacon Transmit Power | 0 dbm Beacon Transmission Interval | 10 Hz Beacon Calibrated Power | Unused Edge Type | Raspberry Pi 3 Raspberry Pi OS | Raspbian Jessie Lite Raspberry Pi Software | Beaconpi Edge Placement Height | 1.6 Metres We use Raspberry Pis and BLE beacons as our receivers and for brevity we subsequently use edge and beacon' to refer to these. The data recording was done via the Beaconpi software. Beaconpi runs both on the edges and a centralized server which communicates bidirectionally between the two main components. The edges and server components sync to the same clock to ensure all elements of the system have a consistent clock. The frame records the fingerprint and the Received Signal Strength Indicator (RSSI) at the edge. To validate our methods we also collect ground truths via an Android application (MapTracker) as the participants move around the experiment area. The Android application records the location of the wearer via self reporting and also collects the accelerometer and gyroscope data for the participant. The participants were asked to randomly walk around the room for around 3 minutes at least 3 times a day. This varied as is evident from the dataset which was expected. The participants walked in straight lines between landmarks that were placed frequently in the space. More details exist in the README.md file. The dataset can also be accessed in GitHub: BLE Beacon Indoor Localization Dataset (BBIL Dataset) [Copyright (c) 2019 Maeve Kennedy]

本数据集旨在推动基于蓝牙低功耗（BLE, Bluetooth Low Energy）信标的室内定位研究。数据采集于2018年9月至2019年5月，覆盖两个独立实验场地。多名受试者参与本实验，每人携带一台BLE信标与一部智能手机。本数据集发布的初衷是为研究者提供室内定位技术的基准测试数据集。我们最初发起该数据集的采集工作，缘于当时带有标注数据的可用室内定位数据集较为匮乏。
本研究采用固定接收端、移动发射端的室内定位方案。我们分别以树莓派（Raspberry Pi）与BLE信标作为接收端与发射端。
安卓最低兼容版本 | 4.4
信标型号 | Gimbal Series 10
信标配置文件 | iBeacon
信标发射功率 | 0 dBm
信标传输间隔 | 10 Hz
信标校准功率 | 未使用
边缘节点型号 | 树莓派3（Raspberry Pi 3）
树莓派操作系统 | Raspbian Jessie Lite
树莓派配套软件 | Beaconpi
边缘节点部署高度 | 1.6米

为行文简洁，后续我们将以“边缘节点”与“信标”分别指代上述接收端与发射端。数据采集通过Beaconpi软件完成，该软件同时部署于边缘节点与中央服务器，二者可实现双向通信。边缘节点与服务器组件同步至同一时钟源，以确保系统所有元素的时钟保持一致。每一条数据帧均记录了接收端处的指纹信息与接收信号强度指示（RSSI, Received Signal Strength Indicator）。

为验证所提方法的有效性，我们还通过一款安卓应用（MapTracker）采集受试者移动时的真实位置标签。该安卓应用可通过受试者自主上报记录佩戴者位置，同时采集受试者的加速度计与陀螺仪数据。

受试者被要求在实验场地内随机行走，每次约3分钟，每日至少进行3次。实验过程存在一定变量，这一点在数据集中亦有体现，符合预期。受试者在场地内频繁设置的地标之间沿直线行进。更多细节可参见README.md文件。本数据集可在GitHub获取：BLE信标室内定位数据集（BBIL数据集）[版权所有 (c) 2019 Maeve Kennedy]

搜集汇总

背景与挑战

背景概述

该数据集是一个用于BLE信标室内定位研究的基准数据集，包含2018年9月至2019年5月在两个地点收集的数据。它采用固定接收器（Raspberry Pi）和移动发射器（BLE信标）的方法，采集了RSSI信号、指纹以及通过Android应用程序获取的地面真实数据（包括位置、加速度计和陀螺仪数据），旨在解决室内定位领域标记数据缺乏的问题。