Effective Fingerprint Extraction and Positioning Method Based on Crowdsourcing

This paper describes an automated method for the calibration of RSS-fingerprinting-based positioning systems. The method assumes using a robotic platform to gather fingerprints in the system environment and using them for training machine learning models. The obtained models are used for positioning purposes during the system operation. The presented calibration method covers all steps of the system calibration, from mapping the system environment using a GraphSLAM based algorithm to training models for radio map calibration. The study analyses four different models: fitting a log-distance path loss model, Gaussian Process Regression, Artificial Neural Network and Random Forest Regression. The proposed method was tested in a BLE-based indoor localisation system set up in a fully furnished apartment. The results have shown that the tested models allow for localisation with accuracy comparable to those reported in the literature. In the case of the Neural Network regression, the median error of robot positioning was 0.87 m. The median of trajectory error in a walking person localisation scenario was 0.4 m

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Automated Calibration of RSS Fingerprinting Based Systems Using a Mobile Robot and Machine Learning

Kołakowski

2021

“…Other technologies that are readily available in smartphone devices used for crowdsensing, such as Bluetooth [9], cellular [10], accelerometers [11], magnetic sensors [12], or pedometers [13], can also be integrated in fingerprinting positioning systems as has been previously shown.…”

Section: Introductionmentioning

confidence: 99%

Crowdsensing Influences and Error Sources in Urban Outdoor Wi-Fi Fingerprinting Positioning

Leca

Nicolaescu

Ciotîrnae

2020

Wi-Fi fingerprinting positioning systems have been deployed for a long time in location-based services for indoor environments. Combining mobile crowdsensing and Wi-Fi fingerprinting systems could reduce the high cost of collecting the necessary data, enabling the deployment of the resulting system for outdoor positioning in areas with dense Wi-Fi coverage. In this paper, we present the results attained in the design and evaluation of an urban fingerprinting positioning system based on crowdsensed Wi-Fi measurements. We first assess the quality of the collected measurements, highlighting the influence of received signal strength on data collection. We then evaluate the proposed system by comparing the influence of the crowdsensed fingerprints on the overall positioning accuracy for different scenarios. This evaluation helps gain valuable insight into the design and deployment of urban Wi-Fi positioning systems while also allowing the proposed system to match GPS-like accuracy in similar conditions.

“…However, these positioning technologies require additional devices, making these systems impractical. Additionally, in signal NLOS propagation, the performance of the positioning systems will be greatly reduced [5].…”

Section: Introductionmentioning

confidence: 99%

“…The experimental results convincingly show that the proposed DNN can automatically learn the location features of LTE signals and achieve satisfactory positioning accuracy in outdoor environments.technologies require additional devices, making these systems impractical. Additionally, in signal NLOS propagation, the performance of the positioning systems will be greatly reduced [5].In contrast, wireless fingerprint positioning technology has received much attention owing to its simplicity and practicality with existing infrastructure and hardware. Many complex LTE signal cues are hidden in the surrounding environment, and the goal of wireless fingerprint-based localization is to discover these cues and use them effectively to determine UE positions [6].…”

mentioning

confidence: 99%

A Novel Outdoor Positioning Technique Using LTE Network Fingerprints

Lei

Zhang

2020

In recent years, wireless-based fingerprint positioning has attracted increasing research attention owing to its position-related features and applications in the Internet of Things (IoT). In this paper, by leveraging long-term evolution (LTE) signals, a novel deep-learning-based fingerprint positioning approach is proposed to solve the problem of outdoor positioning. Considering the outstanding performance of deep learning in image classification, LTE signal measurements are converted into location grayscale images to form a fingerprint database. In order to deal with the instability of LTE signals, prevent the gradient dispersion problem, and increase the robustness of the proposed deep neural network (DNN), the following methods are adopted: First, cross-entropy is used as the loss function of the DNN. Second, the learning rate of the proposed DNN is dynamically adjusted. Third, this paper adopted several data enhancement techniques. To find the best positioning fingerprint and method, three types of fingerprint and five positioning models are compared. Finally, by using a deep residual network (Resnet) and transfer learning, a hierarchical structure training method is proposed. The proposed Resnet is used to train with the united fingerprint image database to obtain a positioning model called a coarse localizer. By using the prior knowledge of the pretrained Resnet, feed-forward neural network (FFNN)-based transfer learning is used to train with the united fingerprint database to obtain a better positioning model, called a fine localizer. The experimental results convincingly show that the proposed DNN can automatically learn the location features of LTE signals and achieve satisfactory positioning accuracy in outdoor environments.technologies require additional devices, making these systems impractical. Additionally, in signal NLOS propagation, the performance of the positioning systems will be greatly reduced [5].In contrast, wireless fingerprint positioning technology has received much attention owing to its simplicity and practicality with existing infrastructure and hardware. Many complex LTE signal cues are hidden in the surrounding environment, and the goal of wireless fingerprint-based localization is to discover these cues and use them effectively to determine UE positions [6]. Compared to satellite navigation positioning systems and other wireless positioning technologies, fingerprint-based positioning techniques have many merits. First, low-energy consumption sensors in UE require much less energy. Second, most of the fingerprint-based positioning technologies require no additional devices or infrastructure. Finally, fingerprint-based positioning can be achieved by using ubiquitous smartphones and signals from LTE base stations (BSs) [6,7]. Fingerprint positioning technology works in two phases: an offline training phase and an online matching phase [8]. During the training phase, signal measurements can be used, which are usually received signal strength indicator (RSSI), reference signal receivin...