PaperlayerView exact revision on arXiv

Source-linked AI summary

A Survey of Indoor Localization Systems and Technologies

Faheem Zafari, Athanasios Gkelias, Kin Leung

arXiv:1709.01015v3cs.NI

TL;DR

Existing surveys lack exhaustive, detailed coverage of indoor-localization techniques, technologies, and recent accurate systems. This paper provides a detailed survey and evaluates proposed systems across multiple practical metrics while identifying remaining challenges.

  • Problem

    Existing surveys do not exhaustively discuss localization access technologies and techniques, and many focus on specific domains.

  • Method

    The paper surveys localization techniques, technologies, and systems, then evaluates systems using accuracy, energy efficiency, scalability, cost, latency, range, and availability.

  • Results

    The survey characterizes proposed indoor-localization systems across practical evaluation metrics and highlights challenges including multipath effects and privacy concerns.

  • Takeaways & Limitations

    Indoor localization supports diverse IoT-enabled services, but accurate and broadly deployable systems must address practical performance and adoption challenges.

  • Takeaways & Limitations

    Wide-scale adoption faces unresolved privacy and security concerns because location data is sensitive and existing systems lack underlying privacy mechanisms.

Abstract

from arXiv · show

Indoor localization has recently witnessed an increase in interest, due to the potential wide range of services it can provide by leveraging Internet of Things (IoT), and ubiquitous connectivity. Different techniques, wireless technologies and mechanisms have been proposed in the literature to provide indoor localization services in order to improve the services provided to the users. However, there is a lack of an up-to-date survey paper that incorporates some of the recently proposed accurate and reliable localization systems. In this paper, we aim to provide a detailed survey of different indoor localization techniques such as Angle of Arrival (AoA), Time of Flight (ToF), Return Time of Flight (RTOF), Received Signal Strength (RSS); based on technologies such as WiFi, Radio Frequency Identification Device (RFID), Ultra Wideband (UWB), Bluetooth and systems that have been proposed in the literature. The paper primarily discusses localization and positioning of human users and their devices. We highlight the strengths of the existing systems proposed in the literature. In contrast with the existing surveys, we also evaluate different systems from the perspective of energy efficiency, availability, cost, reception range, latency, scalability and tracking accuracy. Rather than comparing the technologies or techniques, we compare the localization systems and summarize their working principle. We also discuss remaining challenges to accurate indoor localization.

I. INTRODUCTION

Indoor localization supports services for users and devices across IoT-enabled environments, but existing surveys leave a need for a generic, up-to-date treatment of techniques, technologies, and recent systems. This paper addresses that need through a detailed survey, evaluation framework, IoT primer, and discussion of applications and challenges.

  • Motivation: Indoor localization estimates device or user location indoors and supports applications in health, industry, disaster management, buildings, surveillance, and IoT.The paper distinguishes device-based localization for navigation from monitor-based localization for tracking.
  • Definitions: Device-based localization uses reference or anchor nodes to estimate a user device’s relative location, whereas monitor-based localization passively obtains the user’s position.The two modes have different requirements for energy efficiency, scalability, and performance.
  • Research gap: Existing surveys do not exhaustively discuss access technologies and localization techniques, and many focus on specific domains rather than generic indoor localization.The paper therefore includes tutorial material for readers without prior expertise and considers IoT-related opportunities and challenges.
  • Contributions: The paper surveys user-device tracking systems proposed from 1997 to 2018, emphasizing recent solutions, including some that achieve sub-meter accuracy.It evaluates systems using an evaluation framework intended to highlight their advantages and disadvantages.
  • Contributions: The survey covers localization techniques, technologies, and systems while discussing IoT-induced challenges and applications.Its evaluation framework considers availability, cost, energy efficiency, reception range, tracking accuracy, latency, and scalability.

C. Structure of the Paper

The paper proceeds from localization techniques and wireless technologies to IoT considerations, evaluation metrics, surveyed systems, applications, and challenges. Its RSS-based discussion explains distance estimation, lateration, and accuracy limitations caused by indoor propagation conditions.

  • Paper structure: Section II introduces RSSI, CSI, AoA, ToF, TDoA, RToF, and PoA, plus fingerprinting and machine-learning methods for RSSI-based localization.The discussed learning methods include probabilistic methods, neural networks, kNN, and SVM.
  • Paper structure: Section III discusses WiFi, Bluetooth, Zigbee, RFID, UWB, visible light, acoustic signals, and ultrasound from a localization perspective.The section addresses the advantages and challenges of each technology.
  • Paper structure: The paper introduces IoT localization challenges, evaluation metrics, surveyed systems, applications, and current challenges in Sections IV–VIII.The evaluation framework includes availability, cost, energy efficiency, reception range, tracking accuracy, latency, and scalability.
  • RSS-based localization: RSS-based localization estimates transmitter–receiver distance from received signal strength and uses trilateration or N-point lateration to determine device location.In device-based localization, the device measures RSS relative to at least three reference points; monitor-based localization uses RSS measured at reference points.
  • RSS-based localization: RSS-based localization is simple and cost efficient but can have poor accuracy because walls, obstacles, multipath fading, and indoor noise cause attenuation and signal fluctuations.Filtering or averaging can mitigate these effects, but high accuracy may require more complex approaches.

B. Channel State Information (CSI)

CSI provides richer channel information than RSS by capturing amplitude and phase across frequencies and antenna pairs. The paper relates this information to multipath-aware measurements and fingerprinting or learning-based location estimation.

  • CSI motivation: Wireless channels can be frequency selective, and channel responses may vary across antenna pairs, making RSS susceptible to multipath effects, interference, and temporal variability.These effects arise because different frequencies and antenna pairs can exhibit different amplitude and phase behavior.
  • CSI representation: CSI captures amplitude and phase responses across frequencies and transmitter–receiver antenna pairs, providing finer-grained channel information than RSS.Subcarrier-level measurements from many IEEE 802.11 network cards can provide richer multipath information and more stable measurements.
  • Fingerprinting: Fingerprinting collects offline RSSI or CSI measurements and compares online measurements with stored environmental fingerprints to estimate a user or device position.Fingerprinting maps positions to a discrete grid, so it produces discrete rather than continuous location estimates.
  • Fingerprinting: Probabilistic localization selects the candidate location with the highest likelihood for the observed RSSI vector.With equal prior probabilities, the decision can be based on comparing observation likelihoods across candidate locations; independent reference nodes can be combined multiplicatively.
  • Learning-based localization: Neural networks, kNN, and SVM can use offline and online RSSI measurements to estimate or classify user locations.kNN averages the nearest database matches, while an ANN is trained using RSSI values and corresponding coordinates.

D. Angle of Arrival (AoA)

AoA estimates signal arrival angles with receiver antenna arrays and can localize users with relatively few monitors. The surrounding techniques use propagation time, signal phase, or round-trip timing to infer distances and positions.

  • D. Angle of Arrival (AoA): AoA uses receiver antenna arrays to estimate the angle at which transmitted signals arrive.It exploits time differences of arrival across individual antenna elements.
  • D. Angle of Arrival (AoA): Two monitors can support AoA localization in 2D, while three monitors can support it in 3D.
  • E. Time of Flight (ToF): ToF multiplies signal propagation time by signal velocity to estimate transmitter–receiver distance, then uses geometry for device localization.Three reference nodes are illustrated for estimating distances to the device.
  • E. Time of Flight (ToF): ToF accuracy depends on synchronization, timestamps, bandwidth, sampling rate, and line-of-sight availability in multipath environments.Obstructed direct paths create longer propagation paths and significant localization errors that bandwidth and super-resolution cannot eliminate.
  • F. Time Difference of Arrival (TDoA): TDoA converts propagation-time differences into distance differences and locates the receiver at the intersection of hyperboloids.At least three transmitters are needed for the exact receiver location described here.
  • G. Return Time of Flight (RToF): RToF estimates distance from round-trip signal time and requires less clock synchronization than ToF, but remains affected by sampling rate and bandwidth.

III. TECHNOLOGIES FOR LOCALIZATION

This section surveys radio and wireless technologies used for indoor localization, including WiFi, Bluetooth, Zigbee, RFID, UWB, visible light, and acoustic systems. It outlines their localization mechanisms, capabilities, and practical limitations.

  • III. TECHNOLOGIES FOR LOCALIZATION: The survey presents radio technologies and compares localization techniques through their advantages and disadvantages.
  • A. WiFi: WiFi is widely studied for indoor localization because existing access points can serve as reference points without additional infrastructure.WiFi localization can use RSS, CSI, ToF, AoA, or hybrid methods.
  • A. WiFi: WiFi networks optimized for communication require new algorithms for localization, while uncontrolled ISM-band interference can reduce accuracy.
  • B. Bluetooth: Bluetooth Low Energy supports localization through RSSI, AoA, and ToF, but most systems use RSS-based inputs because they are less complex.RSS-based reliance limits localization accuracy.
  • C. Zigbee: iBeacon systems use periodic BLE messages and RSSI to classify users as immediate, near, far, or unknown relative to a beacon.The reported proximity regions are immediate (<1m), near (1-3m), far (>3m), and unknown.
  • C. Zigbee: iBeacons report only averaged RSSI once per second despite 50 ms transmission intervals, creating challenges for real-time localization.

D. Radio Frequency Identification Device (RFID)

This section contrasts RFID and other wireless or sensing technologies by their operating characteristics, localization accuracy, and deployment constraints. It emphasizes how range, interference, infrastructure, and device limitations shape suitability for indoor localization.

  • D. Radio Frequency Identification Device (RFID): Active RFID supports localization and tracking over reasonable ranges with low cost, but cannot achieve sub-meter accuracy and is not readily available on most devices.
  • D. Radio Frequency Identification Device (RFID): Passive RFID operates without batteries and costs less than active RFID, but its 1-2m range makes it unsuitable for indoor localization.
  • E. Ultra Wideband (UWB): UWB uses sub-nanosecond pulses over bandwidths above 500MHz, reducing power consumption and sensitivity to multipath effects.Its short pulses support accurate ToF estimation, while metals and liquids can interfere with signals.
  • E. Ultra Wideband (UWB): UWB ToF estimation has achieved localization accuracy up to 10cm, although slow standard development has limited adoption in consumer and portable devices.
  • F. Visible Light Communication (VLC): Visible-light localization uses sensors to measure the position and direction of LED emitters, while acoustic systems estimate location from microphone-captured signals.
  • G. Acoustic Signal: Acoustic localization is constrained by smartphone microphone limitations and requires low-power imperceptible signals plus advanced processing.Additional acoustic infrastructure and high update rates can affect battery life and adoption.
  • H. Ultrasound: Ultrasound systems use ToF and sound velocity to provide centimetre-level accuracy, track multiple mobile nodes, and avoid leakage between rooms.

IV. LOCALIZATION AND INTERNET OF THINGS

IoT connects heterogeneous devices to support ubiquitous connectivity and indoor localization, but localization introduces constraints involving privacy, interference, energy, hardware, and protocol diversity. Emerging IoT technologies differ substantially in their suitability for localization.

  • IoT Architecture: IoT connects heterogeneous devices through sensing, communication, and data processing to provide ubiquitous connectivity and enhanced services.Sensors collect data, communication technologies transmit it, and processing infrastructure aggregates and analyzes it.
  • Challenges: Indoor localization in IoT can expose users’ health, mood, behavior, habits, and industrial information, making privacy preservation an open challenge.Data from multiple indoor IoT sensors can reveal behavior even when users do not carry mobile devices.
  • Challenges: IoT ecosystems combine many wireless protocols, whereas most localization systems support only a single air interface, making ubiquitous cross-protocol localization difficult.Modifying device firmware or adding localization sensors is presented as an alternative, but devices may belong to different providers or organizations.
  • Challenges: Localization signaling can interfere with IoT and medical-equipment communications, while frequent updates may impose significant overhead and battery consumption.These effects are especially important for short, infrequent IoT transmissions and devices operating in ISM bands.
  • Challenges: Low-cost IoT devices constrain computation, antenna arrays, signal processing, and battery capacity, limiting the localization techniques they can support.Embedded antennas fit small devices but often provide lower gain and directivity.
  • Emerging Technologies: SigFox and LoRaWAN provide long-range, low-power communication, but long distances, multipath, and limited RSS or bandwidth resolution reduce their localization suitability.SigFox’s ultra-narrowband signal is susceptible to multipath and fast fading, while LoRaWAN’s narrow bandwidth makes multipath resolution difficult.

3) IEEE 802.11ah:

The survey evaluates localization systems using availability, cost, energy efficiency, range, accuracy, latency, and scalability, while emphasizing application-dependent trade-offs. IEEE 802.11ah and other technologies illustrate differing localization capabilities and constraints.

  • Evaluation Framework: Localization systems are evaluated by availability, cost, energy efficiency, reception range, accuracy, latency, and scalability.The paper does not impose universal thresholds because suitable values depend on the application and deployment scale.
  • Availability: User-device availability matters because systems requiring proprietary hardware may limit adoption, despite UWB systems achieving 10-20 cm accuracy.Most current user devices do not include UWB chips.
  • Cost: Localization cost includes infrastructure and user-device requirements, with proprietary reference nodes improving accuracy but increasing expense.The paper notes that smaller businesses are particularly constrained by such costs.
  • Energy Efficiency: Energy consumption should be minimized because localization is an additional service that can drain user-device batteries.BLE and server-side computation are identified as ways to reduce device energy use, with offloading adding latency.
  • Energy Efficiency: Beacon periodicity and transmission power create trade-offs among energy efficiency, accuracy, and reception range.Higher beacon frequency increases energy consumption and accuracy, while higher transmission power increases range and reduces energy efficiency.
  • Energy Efficiency: Algorithmic complexity can drain device power, although offloading computation to a server can reduce device load at the cost of added latency.The paper therefore favors localization algorithms that are not computationally complex on user devices.

D. Reception Range

Indoor localization must balance range, accuracy, latency, and scalability in environments affected by obstacles and multipath. The survey compares device- and monitor-based systems using these application-dependent criteria.

  • Accuracy: Indoor environments challenge localization accuracy through obstacles, multipath effects, and environmental noise.The survey identifies locating users or objects within 10 cm as the ideal microlocalization target.
  • Latency: Real-time localization requires millisecond-scale processing and a small number of reference signals, although additional measurements can improve reliability.The paper frames low latency and reliable estimation as competing requirements.
  • Scalability: Scalable systems must locate many users simultaneously across large spaces such as malls, hospitals, and sports arenas.Monitor-based localization is especially challenged because a server may facilitate hundreds of users at once.
  • Evaluation Framework: No proposed system satisfies all evaluation requirements, although some recently proposed systems satisfy most of them.The framework is intended for application- and deployment-specific comparison rather than universal thresholds.
  • Evaluation Framework: The survey classifies systems as device-based or monitor-based and compares them by energy efficiency, cost, availability, latency, range, accuracy, and scalability.Monitor-based systems are further organized by wireless technology.
  • WiFi-Based Systems: WiFi systems span median accuracies from 2.94 m for RADAR to 23 cm for ArrayTrack, while ToneTrack reports 0.9 m and Phaser 1-2 m.These systems use RSSI, AoA, or ToF with different hardware and signal-processing designs.

2) UWB based MBL:

The surveyed monitor-based systems use UWB, acoustic, RFID, Bluetooth, BLE, ultrasonic, and hybrid signals, exposing trade-offs among accuracy, cost, range, latency, and infrastructure. Reported performance ranges from room-level detection to 10 cm accuracy.

  • UWB-Based MBL: Ubisense achieves accuracy as high as 15 cm and is used commercially, but its cost is a leading constraint.The system is identified as a widely known UWB-based monitor localization system.
  • Acoustic Systems: Beep uses synchronized acoustic signals and distributed sensors connected through WiFi to calculate time of flight for three-dimensional positioning.A user device requests positioning services before transmitting a predefined acoustic signal.
  • Acoustic Systems: BeepBeep performs two-way time-of-flight ranging between commodity devices without proprietary hardware, making it suitable for proximity detection rather than tracking.Both devices emit and record acoustic beeps, then exchange timing information.
  • RFID Systems: LANDMARC tracks users with active RFID tags and reference nodes that measure signal strength and report estimates to a monitor-based server.The reference nodes communicate with the server through IEEE 802.11b WiFi cards.
  • Bluetooth Systems: Bluetooth Location Network provides room-level accuracy with an approximately 11-second response time, making it non-real-time and more suitable for proximity services.Bluetooth reference nodes forward location information to a master node connected to service servers.
  • BLE Systems: iBeacon localization reports average accuracies of 1.441 m with PF, 1.03 m with KF-PF, and 0.95 m with PF-EKF, but requires beacons and incurs delay.The system is described as energy efficient and accurate, while beacon deployment adds cost.
  • Ultrasonic Systems: Cricket achieves 10 cm accuracy using RF for synchronization and ultrasound for localization, but requires dedicated hardware and has limited range.Its design is complementary to the BAT system because receiver-transmitter synchronization is not required.

7) Visible Light based MBL:

The surveyed systems use diverse sensing modalities and signal-processing methods for indoor user localization, with accuracy shaped by environmental and hardware constraints. Acoustic systems offer centimeter-scale accuracy but limited range and deployment suitability.

  • LocaLight uses floor-mounted RFID sensors with photodiodes to detect user shadows without batteries or external power.Under specific LED height, light-zone radius, and individual-height settings, it achieves 50 cm accuracy.
  • The surveyed systems illustrate that localization performance depends on signal conditions, receiver requirements, infrastructure, and deployment environment.These constraints determine whether a system supports exact positioning, proximity estimation, or broader indoor use.
  • UWB systems address indoor non-line-of-sight conditions by modeling signal propagation and classifying LoS versus NLoS signals.Experiments combine extracted signal features with regression analysis and machine learning.
  • Guogou uses inaudible acoustic signals from proprietary reference nodes and signal processing to identify NLoS signals for localization.Its median localization accuracy ranges between 6-25 cm, but proprietary nodes, short range, and noise limit ubiquitous deployment.
  • WalkieLokie supports proximity and relative-position services rather than exact location using acoustic measurements on the user device.The device must remain outside pockets, its range is less than 8 m, and its mean ranging accuracy is 0.63 m.

4) RFID based DBL:

RFID, Bluetooth, visible-light, and emerging IoT systems provide varied indoor localization capabilities, trading accuracy against infrastructure, energy, latency, scalability, and environmental requirements. The survey emphasizes that reported accuracies are not directly comparable across differing experimental settings.

  • RFID based DBL: RFID systems use passive or active tags, readers, RSSI, overlapping signal regions, and movement patterns to estimate mobile-device or user positions.Increasing tag density can improve accuracy but incurs additional cost, while absolute positioning requires known tag locations.
  • BLE based DBL: iBeacon and WiFi fusion improves localization accuracy by 23%, achieving a median accuracy of 0.77 m.The system uses offline RSSI fingerprints and online RSSI measurements, but cannot function in real time and requires multiple values for accurate estimation.
  • Visible Light based DBL: Pharos achieves median accuracy as high as 0.3 m but requires modified LEDs and a proprietary user-device detector, increasing cost and reducing attractiveness.Epsilon provides sub-meter accuracy but requires line of sight and at least three LED reference nodes.
  • Visible Light based DBL: LED TDoA localization achieves an average accuracy of 1.8 mm in simulation over a 75m3 space, subject to a line-of-sight path.The authors do not report energy efficiency or latency, leaving experimental evaluation necessary.
  • Emerging IoT technologies’ based localization: LPWAN positioning remains coarse: OTDOA errors exceed 50 meters for more than 50% of measurements, while LoRa and Sigfox accuracies are in tens of meters.The survey states that current LPWANs require combination with other techniques or technologies for higher accuracy.
  • Miscellaneous Systems: No single localization algorithm is best across conditions, and reported system accuracies cannot be compared one to one because environments and experimental setups vary.Relevant differences include space size, obstacles, number of people, and whether localization is stationary or mobile.

VII. APPLICATIONS OF LOCALIZATION

Indoor localization supports IoT services involving people, objects, and devices across diverse environments. Applications include tracking, navigation, customer assistance, and asset management.

  • Applications: IoT environments such as smart houses, hospitals, schools, malls, and factories create settings for indoor localization services.
  • Tracking: Localization supports both static and mobile-device scenarios, with measurements taken either at fixed points or while users move.
  • Evaluation: The surveyed literature includes existing localization systems and claimed accuracy results for comparing proposed solutions.
  • Technologies: IoT connectivity combines heterogeneous technologies and standards, including BLE, WiFi, Zigbee, RFID, and UWB, for connected localization services.
  • Applications: Indoor localization applications include dedicated sensing for absolute or relative positions and location enrichment for IoT data.These uses can make localization either the primary purpose or an added capability of an IoT system.

A. Contextual Aware Location based Marketing and Customer Assistance

Contextual-aware location-based marketing uses customer location and context to make advertisements and notifications more relevant. Libraries and airports can similarly use localization to support finding items, information, gates, and terminals.

  • Motivation: Traditional advertisements generally do not account for customer location or contextual information such as age, ethnicity, or gender.
  • Marketing: Contextual-aware marketing can target customers with advertisements and coupons relevant to their interests instead of sending irrelevant promotions.
  • Customer Assistance: Libraries can use localization to help visitors find specific books and provide information based on their location.
  • Customer Assistance: Airports can use localization to help customers find boarding gates or terminals without unnecessary inconvenience or wasted time.

B. Health Services

Indoor localization can support health, disaster-response, security, and asset-management services. Its accuracy and adoption remain constrained by environmental dynamics, multipath effects, privacy, energy, cost, and standardization challenges.

  • Health Services: Hospital localization can identify nearby qualified staff during emergencies, reducing reliance on broad messages or paging unavailable personnel.
  • Disaster Management: Disaster-management localization can provide information about whether people are safe and where they are in affected areas.
  • Security: Security and military applications include identifying potential threats, tracking assets and troops, and supporting navigation in unfamiliar areas.
  • Asset Management: Asset tracking can improve inventory and operational management, while energy-efficient techniques may reduce reliance on expensive proprietary hardware.
  • Challenges: Multipath from walls, metals, and people can distort signals used by RSSI, ToF, TDoA, and AoA localization approaches.
  • Challenges: Localization performance varies with indoor materials, obstacles, crowding, and line-of-sight conditions, requiring models that represent environmental variation.

C. Energy Efficiency

Indoor localization adoption requires balancing accuracy, energy use, privacy, cost, interoperability, and latency. The paper identifies standardization, benchmarking, infrastructure-aware mechanisms, and efficient handovers as important directions.

  • Energy Efficiency: Higher accuracy and range commonly require more energy, while periodic beacon listening can strain user-device batteries.
  • Privacy and Security: Privacy is a fundamental adoption challenge because location data is sensitive, while weak authentication can enable malicious attacks.
  • Cost: Additional anchors, infrastructure, servers, databases, and proprietary software can increase the cost of large-scale localization.
  • Standardization: The absence of widely accepted specifications and a dominant wireless technology motivates standardization and universal benchmarking.
  • Infrastructure: Using WiFi access points for localization should avoid jeopardizing their primary connectivity function.
  • Handovers: Heterogeneous networks may require vertical or horizontal handovers, but these must satisfy stringent latency and resource constraints with low-complexity algorithms.
  • Evaluation: The proposed evaluation framework assesses systems using energy efficiency, accuracy, scalability, reception range, cost, latency, and availability.
Loading 1709.01015v3…