Pseudo Label-Driven Federated Learning-Based Decentralized Indoor Localization via Mobile Crowdsourcing

With growing technology, Location Based Services (LBSs) are becoming a vital part of life. The rapid development of mobile internet has offered an opportunity for Wi-Fi indoor positioning to come under the spotlight due to its low cost. Here, indoor positioning is performed using Wi-Fi. The earlier solutions like GPS does not provide efficient positioning in indoor environment due to weak signal reception which makes Wi-Fi more reliable. However, nowadays the accuracy of the traditional indoor positioning algorithms like trilateration cannot meet the practical requirement. This paper proposes the performance analysis of indoor positioning algorithms using Wi-Fi technology based on Received Signal Strength (RSS) from the access points in the network. The positioning algorithms are based on location fingerprinting consisting of two stages: The offline database creation and the online positioning. The RSS values obtained by the mobile device at reference points are sent to the server where all the location fingerprints of all reference points are stored in a database. Coordinates of user position are calculated using positioning algorithms and the position is displayed in terms of 2Dcoordinates of the room and indoor map. Basic Fingerprint algorithm, RSS Fingerprint with minimum Euclidean Distance algorithm and RSS Fingerprint with Euclidean Distance and k-nearest neighbors' algorithm were implemented. Moreover comparison of their performance was carried out in terms of accuracy of position. From the result, it is found that an average accuracy of 0.4962 meters was achieved by RSS Fingerprint with minimum Euclidean distance algorithm.

In location fingerprinting based indoor positioning system, received signal strength (RSS) indications from a set of Wi-Fi access points are used as a unique fingerprint to identify a specific location. However these RSS fingerprints may become outdated when there are unanticipated environmental changes. Re-measuring RSS fingerprints for all locations to maintain an up-to-date RSS database incurs high operational cost, which is impractical in dynamically changed environment. In this paper, we propose a method to estimate the RSS fingerprint of a specific location from a set of neighboring re-measured RSS fingerprints, called “feedbacks”. The proposed method searches for new feedbacks and some necessary old RSS fingerprints in the cut-off area and then applies plane-interpolation to calculate the new RSS fingerprint for a specific location. Based on simulation results, about 5% of re-measured RSS feedbacks are required to satisfy 80% of positioning correctness in the simulated 30×30 m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> area.

Abstract. Indoor localization has attracted the attention of researchers for wide applications in areas like construction, facility management, industries, logistics, and health. The Received Signal Strength (RSS) based fingerprinting method is widely adopted because it has a lower cost over other methods. RSS is a measurement of the power present in the received radio signal. While this fingerprinting method is very popular, there is a significant amount of effort required for collecting fingerprints for indoor space. In this paper, we propose an RSS fingerprinting method using Augmented Reality (AR) that does not rely on an external sensor resulting in ease of use and maintenance. This method uses spatial mapping techniques to help align the floor plan of existing buildings; then, after the alignment, we map local device coordinates to global coordinates. After this process, we partition the space in equally distanced reference points for RSS fingerprint collection. We developed an application for Microsoft HoloLens to align the floor plan and collect fingerprints on reference points. Then we tested collected fingerprints with existing RSS based indoor localization methods for its accuracy and performance.

The indoor pedestrian positioning methods are affected by substantial bias and errors because of the use of cheap microelectromechanical systems (MEMS) devices (e.g., gyroscope and accelerometer) and the users’ movements. Moreover, because radio-frequency (RF) signal values are changed drastically due to multipath fading and obstruction, the performance of RF-based localization systems may deteriorate in practice. To deal with this problem, various indoor localization methods that integrate the positional information gained from received signal strength (RSS) fingerprinting scheme and the motion of the user inferred by dead reckoning (DR) approach via Bayes filters have been suggested to accomplish more accurate localization results indoors. Among the Bayes filters, while the particle filter (PF) can offer the most accurate positioning performance, it may require substantial computation time due to use of many samples (particles) for high positioning accuracy. This paper introduces a pedestrian localization scheme performed on a mobile phone that leverages the RSS fingerprint-based method, dead reckoning (DR), and improved PF called a double-stacked particle filter (DSPF) in indoor environments. As a key element of our system, the DSPF algorithm is employed to correct the position of the user by fusing noisy location data gained by the RSS fingerprinting and DR schemes. By estimating the position of the user through the proposal distribution and target distribution obtained from multiple measurements, the DSPF method can offer better localization results compared to the Kalman filtering-based methods, and it can achieve competitive localization accuracy compared with PF while offering higher computational efficiency than PF. Experimental results demonstrate that the DSPF algorithm can achieve accurate and reliable localization with higher efficiency in computational cost compared with PF in indoor environments.

LoRa technology has received a lot of attention in the last few years. Numerous success stories about using LoRa technology for the Internet of Things in various implementations. Several studies have found that the use of LoRa technology has the opportunity to be implemented in indoor-based applications. LoRa technology is found more stable and is more resilient to environmental changes. Environmental change of the indoor is a major problem to maintain accuracy in position prediction, especially in the use of Received Signal Strength (RSS) fingerprints as a reference database. The variety of approaches to solving accuracy problems continues to improve as the need for indoor localization applications increases. Deep learning approaches as a solution for the use of fingerprints in indoor localization have been carried out in several studies with various novelties offered. Let’s introduce a combination of the use of LoRa technology's excellence with a deep learning method that uses all variations of measurement results of RSS values at each position as a natural feature of the indoor condition as a fingerprint. All of these features are used for training in-deep learning methods. It is DeepFi-LoRaIn which illustrates a new technique for using the fingerprint data of the LoRa device's RSS device on indoor localization using deep learning methods. This method is used to find out how accurate the model produced by the training process is to predict the position in a dynamic environment. The scenario used to evaluate the model is by giving interference to the RSS value received at each anchor node. The model produced through training was found to have good accuracy in predicting the position even in conditions of interference with several anchor nodes. Based on the test results, DeepFi-LoRaIn Technique can be a solution to cope with changing environmental conditions in indoor localization

Recently, machine learning (ML) has been widely adopted for fingerprint-based indoor localization because of its potency in delineating relationships between received signal strength (RSS) information and labels accurately. Existing ML-based indoor localization systems are less robust because they only adopt the output with the highest probability. This affects the final location estimate, hence compromising accuracy due to the severity of RSS fluctuations. Since different ML algorithms (MLAs) yield different performances, it is therefore intuitive to fuse predictions from multiple MLAs to improve the positioning performance in the presence of signal fluctuation. In this article, we propose SmartLoc, a smart wireless indoor localization framework to enhance indoor localization. In the offline phase, multiple MLAs are trained by utilizing an offline database. We further apply probability alignment to guarantee the predicted probabilities of each MLA at the same confidence level. In the online phase, given a testing RSS sample of a user at an unknown location, we extract the labels with probabilities greater than a certain threshold from each MLA to construct the space of candidate labels (SCL). The size of SCL can be adaptively determined by using our proposed dynamic size determination algorithm. Based on the SCL, we propose a probabilistic model to intelligently estimate the user's location by evaluating the label credibility simultaneously. A high label credibility indicates that the frequently occurred label is more likely to be true. Experimental results in a real changing environment verify the superiority of SmartLoc, outperforming the best among comparative methods by 10.8% in 75th percentile accuracy.

The Internet of Things (IoT) is an emerging paradigm to integrate the physical world into cyber systems by connecting various devices, which sense and control the surrounding objects. For the upcoming IoT era, narrow-band signals will play an important role in the exchange of sensor data and control signals in IoT networks. Indoor positioning is critical for many IoT applications, such as asset tracking, smart buildings, smart e-health, and smart cities, but it is very challenging to achieve accurate indoor positioning with narrowband signals because of multipath propagation. In this paper, we design a positioning system using narrowband signals, particularly ZigBee signals, based on an enhanced fingerprinting algorithm by fusing received signal strength (RSS) and time information. We first investigate the feasibility of time-based fingerprinting by handling the challenges as synchronization compensation among anchor nodes and design of sub-sample timestamps with a resolution of nanoseconds. Second, we design a feature-based fusion approach to fuse and standardize time and RSS fingerprints. Third, we adopt a random forest regression model to design an enhanced pattern matching algorithm for fingerprinting. We implement the proposed algorithms with software-defined radio techniques. In the given experiment results, we find that the positioning system achieves a mean positioning accuracy of 1.61 m, which represents a 36.1% improvement over traditional RSS-based fingerprinting.

Received signal strength (RSS) measurements are easily available and provide context information to IoT-devices. However, indoor localization via RSS is error prone. Due to walls and ceilings multipath propagation occurs and leads to constructive and destructive interference and therefore higher or lower RSS values in an unpredictable and therefore random fashion. This random variation decreases localization accuracy if not considered right. In our work, we develop a pathloss model via a ray tracing engine and rate several measured RSS values by statistical means. We find a good estimate of the RSS of the line of sight connection which improves the distance estimation significantly. Our algorithm is based on the assumption that several frequencies are used for localization, which is true for Bluetooth Low Energy (BLE). We demonstrate the improved RSS-based-localization by real experiments and BLE hardware.

In this paper, we propose an indoor positioning system based on Bluetooth low energy (BLE), and the localization tag is with characteristics of low power, low cost, and high portability. The proposed system consists of BLE tags, BLE/WiFi repeaters, and a fusion server. The BLE tag is a device which broadcasts BLE beacons. The BLE/WiFi repeaters collect the beacons transmitted from the tag and extract the received signal strength (RSS). The RSS values are then transmitted to the fusion server through WiFi, and the server will estimate the position of the BLE tag with RSS-based positioning algorithm. We propose an indoor positioning algorithm which is a hybrid from received signal strength indication (RSSI)-fingerprint and cell of origin (CoO) to estimate the BLE tag's location. Some modifications are made to typical RSSI-fingerprint and CoO algorithm to get better accuracy. The test results show 1.2 m to 1.4 m positioning accuracy. The size of the BLE tag is 1.7 cm in radius and 0.5 cm thick. Each BLE tag costs 3 US dollars. The current consumption of the tag is 50 μA which can be used without charge for 136 days with a CR2025 battery.

Indoor localization schemes have significant potential for use in location-based services in areas such as smart factories, mixed reality, and indoor navigation. In particular, received signal strength (RSS)-based fingerprinting is used widely, given its simplicity and low hardware requirements. However, most studies tend to focus on estimating the 2D position of the target. Moreover, it is known that the fingerprinting scheme is computationally costly, and its positioning accuracy is readily affected by random fluctuations in the RSS values caused by fading and the multipath effect. We propose an indoor 3D localization scheme based on both fingerprinting and a 1D convolutional neural network (CNN). Instead of using the conventional fingerprint matching method, we transform the 3D positioning problem into a classification problem and use the 1D CNN model with the RSS time-series data from Bluetooth low-energy beacons for classification. By using the 1D CNN with the time-series data from multiple beacons, the inherent drawback of RSS-based fingerprinting, namely, its susceptibility to noise and randomness, is overcome, resulting in enhanced positioning accuracy. To evaluate the proposed scheme, we developed a 3D positioning system and performed comprehensive tests, whose results confirmed that the scheme significantly outperforms the conventional common spatial pattern classification algorithm.

By taking advantage of Deep Learning (DL), Received signal strength (RSS) fingerprint-based indoor localization has attracted more attention. Training DL models require an immense amount of RSS samples, and crowdsourcing has been another way to collect data by users. Recently, some researchers apply Federated Learning (FL) in indoor localization instead of crowdsourcing for privacy protection. However, more practical issues are still not considered.For further guaranteeing privacy and deploying systems in real time, this paper proposes OPFL, a privacy-preserved online personalized federated learning framework for fingerprint-based indoor localization. In OPFL, users update model parameters and collect new labeled samples meanwhile. By adding artificial noise based on the Differential Privacy (DP) mechanism, it is harder for attackers to infer information. Besides, the deterioration of models results from DP is mitigating by applying the personalized algorithm. The experimental results show that under the noise perturbing, the performance of OPFL testing on the local test set is far better than the conventional federated averaging algorithm (FedAvg), and is even close to the centralized training.

Context-aware applications have been gaining huge interest in the last few years. With cell phones becoming ubiquitous computing devices, cell phone localization has become an important research problem. Database Correlation Method (DCM) is a positioning technology that based on a database of a premeasured location dependent variable such as Received Signal Strength (RSS). DCM has shown superior in terms of accuracy. Absolute RSS values received from a base station change with time, but the relative RSS (RRSS) values which refer to the relations of the RSS values between different base stations are more stable. This study proposes and implements a robust RRSS GSM-based technique for both positioning and traffic estimation. The study was tested and analyzed in Egypt roads using realistic data and Android smart phones. The performance evaluation showed good results. Mean positioning accuracy was about 29m in urban areas and velocity estimation was about 1 km/h in rural areas.

In indoor location environment, shadowing, multipath and reflection phenomena make the position estimation inaccurate. Large number of Access Points (APs) equipped with single antenna could obtain enough Receive Signal Strength (RSS) information to describe signal characteristics of the target space. However, it results in a large investment costs. This paper provides a novel indoor positioning solution using the multi-antennas architecture to increase the accuracy of location estimation. The system obtains enough RSS to compute the user location and provides a reasonable cost of deployments in indoor location environment. Experimental and numerical results show that the proposed indoor location estimation demonstrates a high accuracy of positioning and outstanding performance when RSS is collected by smart antenna technique. More specifically, the accuracy is increased from 1.34% to 92.4% and the error variance is reduced by 96% as compared to the traditional location fingerprint method.

Context-aware applications have been gaining huge interest in the last few years. With cell phones becoming ubiquitous computing devices, cell phone localization has become an important research problem, and numerous localization solutions have been proposed. Database Correlation Method (DCM) is a positioning technology that based on a premeasured database of a location dependent variable such as Received Signal Strength (RSS) or Link Quality Identifier. DCM has shown superior in terms of accuracy. Absolute RSS values received from base station change with time, but the relative RSS values which refer to the relations of the RSS values between different base stations are more stable. This study has proposed and implemented a robust technique for both positioning and traffic estimation using Relative Received Signal Strength (RRSS) of GSM. The study has been tested and analysed in Egypt roads using realistic data and Android smart phones. The performance evaluation showed good results. Mean positioning accuracy was about 29 m in urban areas and velocity estimation was about 1 km/h in rural areas.

With the popularity of smart mobile devices, “context-aware” applications have attracted intense interest, for which location is one of the most essential contexts. Compared with outdoor localization, indoor localization has received much more attention from both academia and industry these days. Given the widespread use of WiFi hotspots, the received signal strength (RSS) fingerprint-based indoor localization technique is considered as a promising and practical solution because of its relatively high accuracy and low infrastructure cost. Inspired by our observation that sparsity is inherent to the WiFi signal, we present a new RSS fingerprint-based indoor localization approach, called SparseLoc . Through sparse representation of the fingerprints, SparseLoc can estimate a smart mobile device’s location with a small error most of the time. Although the correlation between neighboring fingerprints affects the localization accuracy, SparseLoc uses the similarity between principal components of fingerprints to alleviate this effect. Based on the empirical experiments, we demonstrate that SparseLoc improves the localization accuracy by over 25% compared with the existing WiFi signal-based localization methods.