Marzieh Dashti

Indoor localization using smartphone inertial sensors

Celebrated fingerprinting techniques localize users by statistically learning the signal to location relations. However, collecting a lot of labelled data to train an accurate localization model is expensive and labour-intensive. In this paper, an economic and easy-to-deploy indoor localization model suitable for ubiquitous smartphone platforms is established. The method processes embedded inertial sensors readings through a inertial localization system. A particle filter is developed to integrate the building map constraints and inertial localization results to estimate users location. To increase the algorithm convergence rate, the users initial/on-line room-level localization is achieved using WiFi signals. To achieve room-level accuracy, only very few training WiFi data, i.e. one per room or per segment of a corridor, are required. A novel crowdsourcing technique to build and update training database is presented. On these basis, an indoor localization system is proposed and evaluated. The results show that comparable location accuracy to previous approaches without even dense wireless site survey requirements is achievable.

Indoor localization on mobile phone platforms using embedded inertial sensors

Indoor localization techniques are commonly based on measuring the Wi-Fi received signal strength (RSS) and the method of “fingerprinting”. The accuracy depends on density of recorded fingerprints in the radiomap (RM) database. Building and updating RM is expensive and labor-intensive. Taking advantage of ubiquity of smartphones with embedded inertial sensors yields an economic and easy-to-deploy indoor localization system. In this work, the accelerometer and digital compass are used to recognize the users dynamic activities and walking directions. And the particle filter integrates the building map constraints and inertial measurements to estimate users location. On this basis, an indoor localization system with no dense wireless site survey requirements is proposed and evaluated.

Detecting human encounters from WiFi radio signals

We present the design, implementation and evaluation of a novel human encounter detection framework for measuring and analysing human behaviour in social settings. We propose the use of WiFi probes, management frames of WiFi, that periodically radiate from mobile devices (as proxies for humans), and existing WiFi access points to automatically capture radio signals and detect human copresence. Based on the spatio-temporal properties of this copresence and their interplay we defined a model, borrowing theories from sociology, to detect human encounters -- short-lived, spontaneous human interactions. We evaluated our framework using controlled and in-the-wild experiments yielding a detection performance of 96% and 86% respectively. As such, our framework opens up interesting opportunities for designing proxemic and group applications, as well as conducting large-scale studies in the areas of computational social sciences.

Wireless RSSI fingerprinting localization

Localization has attracted a lot of research effort in the last decade due to the explosion of location based service (LBS). In particular, wireless fingerprinting localization has received much attention due to its simplicity and compatibility with existing hardware. In this work, we take a closer look at the underlying aspects of wireless fingerprinting localization. First, we review the various methods to create a radiomap. In particular, we look at the traditional fingerprinting method which is based purely on measurements, the parametric pathloss regression model and the non-parametric Gaussian Process (GP) regression model. Then, based on these three methods and measurements from a real world deployment, the various aspects such as the density of access points (APs) and impact of an outdated signature map which affect the performance of fingerprinting localization are examined. At the end of the paper, the audiences should have a better understanding of what to expect from fingerprinting localization in a real world deployment.

Detecting co-located mobile users

Co-location information of devices, people, and activities can be used in numerous applications in areas of social networking, mobile networking, spatial and socio-economics, and securing interactions. People co-location can be used to infer their communications and interactions. This information can be exploited for many purposes such as gaining understanding of human social interactions and behaviours. In this paper, we propose a real-time co-localization technique which provides accurate people co-location information with sub-meter accuracy. We construct a connectivity graph representing the potential colocated users based on pairwise similarity of RF measurements from users mobile phones. We then apply community-detection tools to cluster users into co-located groups. Since our approach does not estimate the absolute location of individual users, it is robust to localization errors and protects the location privacy of mobile users. Our approach does not involve labour-intensive calibration as required for most localization approaches. We prototyped our proposed solution to detect co-located users in an enterprise building scenario. Android mobile users connected to our cloud localization server were accurately clustered according to their geographical proximity.

RSSI Localization with Gaussian Processes and Tracking

Location Fingerprinting (LF) is a promising localization technique that enables many commercial and emergency location-based services (LBS). While significant efforts have been invested in enhancing LF using advanced machine learning methods, the configuration effort required to deploy a LF system remains a significant issue. In this paper, a practical LF system is proposed which employs Gaussian Processes (GP) to significantly reduce the required database density. The GP solution is enhanced with a tracking algorithm which can easily incorporate floor plan constraints. The proposed system was prototyped with Android mobile phones in an enterprise environment. It is shown that with the proposed system an accuracy required for most commercial LBS applications can be achieved with a significantly reduced configuration effort.

Localization of unknown indoor wireless transmitter

Many management tasks, for instance optimizing placement of a new Wi-Fi or femtocell access points (AP), or detecting unauthorized transmitter (Tx), requires the ability to locate individual Tx inside buildings. Available techniques to locate Txs require extensive war driving measurements and significant computations, or complex and additional hardware. This paper presents a time-efficient method, based on only collected received signal strength (RSS) data to estimate the location of unknown Tx installed inside a multi-storey building. Three-stage algorithm is proposed. Firstly, the buildings location from which the signal is transmitted from, is defined on the area map. Secondly, the floor level of the determined building on which the unknown Tx is installed, is determined. Finally, 2-dimensional location coordinates of Tx and the path loss parameters are jointly estimated. The method is evaluated using realistic simulated data obtained from iBuildNet® wireless network design and optimization tool developed by Ranplan. The simulation results confirm that developed algorithm works accurately and is especially helpful to locate an unknown Tx in changing and unknown environments.

Locating user equipments and access points using RSSI fingerprints: A Gaussian process approach

Location fingerprinting (LF) is an attractive localization technique which relies on existing infrastructures. The major drawback of LF is the requirement of having an updated fingerprint database. Gaussian Process (GP) is a non-parametric modeling technique which can be used to model the received signal strength indicator (RSSI) and create the fingerprint database based on few training data. In this paper we use a parametric pathloss model for the GP mean and a flexible non-parametric covariance function, so we can get reliable estimates with low fingerprinting effort. In our experiment, we show that with 23 fingerprint locations we perform as well as traditional fingerprinting with over 230 fingerprinted locations for an office space of 2500m2.

Floor determination for positioning in multi-story building

WiFi access points (APs) are nowadays ubiquitous in multi-level buildings to provide uninterrupted network access to mobile users. WiFi signals are the widely recognized option to provide feasible indoor positioning system (IPS). However majority of IP research concentrates on 2D positioning compared to vertical or floor level determination. To date the most reliable floor determination techniques based on WiFi signals use fingerprinting approaches. However, fingerprint (FP) method is computationally extensive due to its large database size. This paper proposed a new computationally efficient floor determination system based on reduced database. The performance of proposed algorithm is tested using real-world measurement and is compared against other few available floor determination algorithms. The experiment shows that the proposed method outperforms other floor positioning methods by utilizing only 14% of original FP database size.

A Survey on Wireless Transmitter Localization Using Signal Strength Measurements

Knowledge of deployed transmittersź (Tx) locations in a wireless network improves many aspects of network management. Operators and building administrators are interested in locating unknown Txs for optimizing new Tx placement, detecting and removing unauthorized Txs, selecting the nearest Tx to offload traffic onto it, and constructing radio maps for indoor and outdoor navigation. This survey provides a comprehensive review of existing algorithms that estimate the location of a wireless Tx given a set of observations with the received signal strength indication. Algorithms that require the observations to be location-tagged are suitable for outdoor mapping or small-scale indoor mapping, while algorithms that allow most observations to be unlocated trade off some accuracy to enable large-scale crowdsourcing. This article presents empirical evaluation of the algorithms using numerical simulations and real-world Bluetooth Low Energy data.