Earticle

지자기 벡터는 센서가 바라보고 있는 방향에 따라 그 값이 달라지는 특성이 있다. 본 논문에서는 그런 문제를 최소 화하여 지자기 기반 실내 위치 추정에 사용될 수 있도록 지자기 벡터 보정법을 제안한다. 지자기 기반 실내 위치 추 정에서 사용되는 핑거프린팅 기법은 자기장 지도와 현재 위치에서의 자기장 값을 매칭하여 위치를 추정해낸다. 이 때, 자기장 센서는 사용자의 이동 방향에 따라 읽어 들이는 자기장 벡터 값이 달라지기 때문에 위치 추정 정확도가 낮아진다. 이를 해결하기 위해 많은 연구들은 자기장 벡터 크기를 사용하지만, 이는 지문의 고유성을 감소시킨다. 따라서 본 논문에서는 지문의 고유성을 유지할 수 있는 자기장 벡터를 그대로 사용하되, 벡터 크기처럼 사용자의 이 동방향에 영향을 받지 않도록 벡터 값을 보정하는 방법을 제안한다. 임의의 방향으로 걸어본 결과, 본 연구에서 제 안된 보정법을 사용하면 자기장 지도와의 매칭 정확도가 높아지는 것을 확인하였다.

Magnetic sensors have the disadvantage that their vector values differ depending on the direction. In this paper, we propose a magnetic vector calibration method for geomagnetic-based indoor localization estimates. The fingerprinting technique used in geomagnetic-based indoor localization the position by matching the magnetic field map and the magnetic sensor value. However, since the moving direction of the current user may be different from the moving direction of the person who creates the magnetic field map at the collection time, the sampled magnetic vector may have different values from the vector values recorded in the field map. This may substantially lower the positioning accuracy. To avoid this problem, the existing studies use only the magnitude of magnetic vector, but this reduces the uniqueness of the fingerprint, which may also degrade the positioning accuracy. In this paper we propose a vector calibration algorithm which can adjust the sampled magnetic vector values to the vector direction of the magnetic field map by using the parametric equation of a circle. This can minimize the inaccuracy caused by the direction mismatch.

근래의 위치 측위 방법으로 GPS(Global Positioning System) 위성정보를 활용하는 전파항법 방 식을 많이 사용하고 있다. GPS 활용범위가 넓어지고 다양한 측위 정보를 기반으로 하는 분야가 생기면서 보다 높은 정확도를 얻기 위한 새로운 방법들이 요구되고 있다. 자율주행차의 경우 IMU(Inertial Measurement Unit)를 사용한 항법 시스템인 INS(Inertial Navigation System)와 차량 내 부 센서를 이용한 DR(Dead Reckoning) 알고리즘을 사용하여 GPS의 정확도 저하나 음영지역에서 의 위치 측정방법으로 사용하고 있다. 그러나 이러한 측위 방법은 대형화되는 빌딩 지역, 터널, 지하 주차장 등 다양한 음영지역과 시간이 지남에 따라 오차가 계속 증가하는 누적 기반 위치추 정 방법의 한계로 인해 많은 문제 요소가 있다. 본 논문은 GPS 음영지역에서 차량의 위치 측위를 위해, 대중적 무선 통신인 WLAN을 이용한 Fingerprint 기법을 4개의 Anchor 형태로 AP(Access Point)와 지향성 안테나를 위치하여 넓은 지하 주차공간에서 효율적인 측위 방법을 제시하고 시간 이 지남에 따라 주차된 차량이 이동하는 환경에서도 변화가 없는 위치 측위 결과를 입증하였다.

Recently, the radio navigation method utilizing the GPS(Global Positioning System) satellite information is widely used as the method to measure the position of objects. As GPS applications become wider and fields based on various positioning information emerge, new methods for achieving higher accuracy are required. In the case of autonomous vehicles, the INS(Inertial Navigation System) using the IMU(Inertial Measurement Unit), and the DR(Dead Reckoning) algorithm using the in-vehicle sensor, are used for the purpose of preventing degradation of accuracy of the GPS and to measure the position in the shadow area. However, these positioning methods have many elements of problems due not only to the existence of various shaded areas such as building areas that are continually enlarged, tunnels, underground parking lots and but also to the limitations of accumulation-based location estimation methods that increase in error over time. In this paper, an efficient positioning method in a large underground parking space using Fingerprint method is proposed by placing the AP(Access Points) and directional antennas in the form of four anchors using WLAN, a popular means of wireless communication, for positioning the vehicle in the GPS shadow area. The proposed method is proved to be able to produce unchanged positioning results even in an environment where parked vehicles are moved as time passes.

위치 추정은 사용자의 위치에 따른 정보와 오락을 제공하는 위치기반 서비스의 가장 중요한 핵심 기술이다. 실내 환경에서의 위치추정은 주행로봇분야에서 매우 활발하게 연구되어왔으며, 가장 널리 쓰이는 방법 중의 하나는 인공 표식을 이용하는 것이다. 그러나 일반 보행자뿐만 아니라 노약자 및 장애인의 움직임은 로봇의 주행과 다르므로, 주행로봇을 위한 위치추정 기술은 보행자에게 그대로 사용될 수 없다. 이 논문에서는 인공표식과 관성항법 측정 장치를 사용한 보행자를 위한 실내 위치 추정 연구에 대하여 기술한다. 인공표식은 실내조명 환경 변화에 강인하며, 적외선이 사용자에게 보이지 않아 시각적인 방해를 최소화하기 위해 적외선 반사체를 이용하였다. 또한 설치해야 하는 인공표식의 수를 줄이기 위하여, 인공표식 기반의 위치 추정기술과 인공표식이 보이지 않는 동안 사용하는 관 성항법 장치 기반의 위치 추정 기술을 동시에 사용하는 융합 위치 추정 방법을 제시하였다.

Abstract Localization is the key technology for location-based service; an information and entertainment service utilizing the ability to make use of the geographical position of the user. Indoor localization is developed widely for mobile robots. One of the localization methods is using the artificial landmarks; however, it can not be applied without modification to the localization of the pedestrian because the pedestrian may walk differently from the mobile robot especially when we are interested in localizing the elderly and the handicapped. This paper presents newly developed pedestrian localization system using the artificial landmarks and inertial measurement unit. Infrared light reflecting landmarks are designed for robustness and the infrared light does not disturb the user. In order to reduce the number of landmarks installed, the inertial measurement unit based deadreckoning is used while the camera doesn't not have the view of the landmark.

With the popularization of sensor technology, researchers extract Channel State Information (CSI) from WiFi, which reflects the movement of family members by observing changes in signal transmission in residences. However, WiFi signals are still affected by multipath effects in residences. Therefore, we construct a multi-scale CSI combination mask matrix between adjacent scales based on the ternary closure. Next, we propose a competitive localization network based on aggregated affinity propagation algorithm. Experiments have been conducted to demonstrate that the proposed algorithm achieves significant improvement compared to other classical algorithms in the indoor environment.

This paper presents a tutorial on Deep Learning (DL) with Federated Learning (FL)-based indoor localization method for non-Independently and Identically Distributed (non-IID) fingerprinting databases. To this end, this paper explains systematic approaches for addressing privacy concerns and performance degradation issues in non-IID fingerprinting databases. The method presented in this tutorial entails the application of a personalized layer, model reliability, and Layer-wise local model’s Weight Change (LWC) information to FL. This tutorial provides intuitions to be considered by future researchers to improve the performance of FL-based fingerprinting localization by summarizing the above-mentioned methods into three FL-based techniques: high-complexity training for performance improvement of local training models, exact characteristics of the local model for global model aggregation, and Bayesian data fusion for probabilistic clustering, to improve FL-based indoor localization performance.

Bluetooth Low Energy (BLE) based indoor positioning systems rely on accurately classifying channel conditions such as line-of-sight (LOS) or non-line-of-sight (NLOS). However, classification models trained in one building rarely generalize to another due to different floor layouts, anchor deployment, and interference patterns. The existing solutions often assume rich channel features, require labels from each new environment, or depend on fixed anchor layouts, which limit their scalability. We propose a BLE based domain adaptive RF channel classification that incorporates adversarial domain alignment and confidence-based pseudo-labeling to leverage unlabeled target data. We evaluate the approach using BLE Received Signal Strength Indicator (RSSI) data collected from three indoor areas: a corridor (source domain), a classroom (target domain), and an office room (unseen test domain). The proposed approach shows 2% gain over the no adaptive classification framework.

The individual markets for indoor localization service (ILS) and augmented reality (AR) are expected to grow significantly. These combined services will be very diverse, and the market is expected to grow significantly. This paper presents representative technologies supporting ILS based on wireless communications and computer visions. Then, AR-enabled ILS cases are also presented.

Internet of Things (Internet of Things, IoT) is becoming commonplace at a faster rate, thus the study of low-power long-distance communication is being accelerated. In addition, localization based on wireless sensor network bas been interested. Positioning using LoRaWAN boasts relatively excellent accuracy compared to other technologies, but it seems a lot of error in the nature of other wireless technologies in indoor environment. This paper correct TDOA localization result using in advance constructed TDOA correction database in such indoor positioning.

In this study, we present an algorithm for indoor robot position estimation. Estimating the position of an indoor robot using a fixed imaging device obviates the need for complex sensors or hardware, enabling easy estimation of absolute position through marker recognition. However, location estimation becomes impossible when the device moves away from the surrounding obstacles or the screen, presenting a significant drawback. To solve this problem, we propose an algorithm that improves the precision of robot indoor location estimation using a Gaussian Mixture Model(GMM) and a Kalman filter estimation model. We conducted an actual robot operation experiment and confirmed accurate position estimation, even when the robot was out of the image.

To locate an RF device or to enable target oriented communication services in wireless indoor environment, RF properties as RSSI and CSI have been widely adopted in the research and commercial societies. However, the complex indoor scenario makes the task to locate an RF device with fine-grained accuracy more difficult. This paper deals with a virtual triangulation utilization to estimate the length of Line of Sight (LoS) signal out of non LoS on receiving multipath signals from any dimension. Thus, it forms a virtual right angled triangle of the received signals. From the angles ratio, the propose method estimates the side corresponding to the reference LoS length. Eventually, it estimates the distance error between the measured reference LoS of the virtual triangle and the approximated LoS from the CSI. The error is utilized to provide a basement for more precise localization of an RF device.

This study presents the results of designing a system that determines the location of a person in an indoor environment based on a single IMU sensor attached to the tip of a person's shoe in an area where GPS signals are inaccessible. By adjusting for human footfall, it is possible to accurately determine human location and trajectory by correcting errors originating from the Inertial Measurement Unit (IMU) combined with advanced machine learning algorithms. Although there are various techniques to identify stepping, our study successfully recognized stepping with 98.7% accuracy using an artificial intelligence model known as Long Short-Term Memory (LSTM). Drawing upon the enhancements in our methodology, this article demonstrates a novel technique for generating a 200-meter trajectory, achieving a level of precision marked by a 2.1% error margin. Indoor pedestrian navigation systems, relying on inertial measurement units attached to the feet, have shown encouraging outcomes.

With the rapid development of wireless communication and embedded system, wireless positioning systems are paid more and more attention to. Radio Frequency Identification (RFID) localization system is getting more important, due to its own advantages, such as no contact, non-line-of-sight nature, promising transmission range and cost-effectiveness. To improve the accuracy of active RFID indoor location system, some traditional RFID indoor localization systems were studied, such as LANDMARC. On this basis, an adaptive self-correction location algorithm was presented, which uses a positioning correction value to correct the positioning result. N minimum errors and position results are obtained by using adaptive K-nearest neighbor algorithm N times. The positioning correction value calculated with N minimum errors in weighted way. The sum of the positioning average value and the positioning correction value would be the final positioning results. Experimental results show that compared with adaptive K-nearest neighbor algorithm and error self-correction algorithm, the proposed method provides a higher accuracy and stability.

In this paper, a localization system for indoor objects is proposed. The proposed system consists of Beacons, LED Cells, Main Cell Controller (MCC), and Display. A Beacon is attached at each indoor object, and each LED cell has Beacon Scanner and VLC Transmitter. The Visual Light Communications (VLC) and Power Line Communications (PLC) methods are used to communicate the signals for localization of indoor objects. And the proposed system is designed, and implemented as a prototype. To certify that our proposed system can exactly localize a given indoor object, we take test for the implemented system as a prototype. Here the location of the given indoor object is known. Test is done in two ways. The first is to check the operation of the detail of the system, and the second is the position recognition of indoor object. The second is the test of the implemented system to correctly detect the location of the indoor object with Beacon, while the object with Beacon is moved from location C to A. The experimental result shows that the system is exactly detect the moving locations. The system has the advantages of using previously installed power lines, and it does not need to use LAN lines or optical cables. The proposed system is usefully applied to indoor object localization area.