Fundamentals of Positioning Technologies: Satellite, Cellular, Wi‑Fi, Bluetooth, Geomagnetic, Inertial, LED, Vision, UWB, and IP

Preface: Quoting a line from "Above the Map: Tracing the Original World" – "Books about maps are actually books about the progress of the world" – maps are humanity’s way of organizing and projecting its understanding of the external world. A cave‑forest map reveals the life of primitive hunters; a city aerial map reveals past civilization; a maritime chart reveals the courage of explorers.

In modern times, maps have evolved into the digital maps on our phones. The term "positioning" frequently appears alongside maps, meaning the determination of a point on a map. In the mobile‑Internet era, digital maps have grown dramatically, and positioning technologies have seen widespread deployment and technical upgrades.

This article will explain, in plain language, the popular positioning principles that let mobile devices determine their location on a map.

Basic Principles

1. Satellite Positioning

1.1 Overview

Satellite positioning calculates location by receiving satellite signals. Main systems: United States (GPS), Russia (GLONASS), Europe (GALILEO), China (BeiDou). GPS is the most widely used and mature.

GPS system three parts: space segment, ground control, user equipment

Space segment : 24 satellites, evenly distributed in 6 orbital planes (4 per plane).

Control segment : monitoring stations, master control stations, and injection stations manage the GPS satellites.

User equipment : GPS receiver (chip) receives signals, obtains measurement values, processes data, and produces a position.

1.2 Satellite Positioning Principle

GPS positioning principle: use four satellites with known positions to determine the receiver’s location.

1) Satellite position: derived from the satellite‑borne clock and ephemeris data.

2) User‑to‑satellite distance: calculated from signal travel time multiplied by the speed of light (the measured distance is a pseudorange because atmospheric and ionospheric delays affect it).

Note: GPS satellites continuously broadcast navigation messages containing ephemeris, health status, clock corrections, ionospheric delay corrections, and atmospheric refraction corrections.

3) Position solving: besides the three unknown coordinates (x, y, z), an additional unknown time offset t is introduced (the receiver’s clock is not synchronized). Four equations solve the four unknowns.

1.3 Multipath Effect

During signal propagation, reflections from objects change the signal’s direction and amplitude before reaching the GPS chip. This distortion, called the multipath effect, is a major source of GPS positioning error.

1.4 A‑GPS

GPS suffers a noticeable cold‑start delay: the receiver must search all 24 satellites, taking about two minutes.

A‑GPS accelerates this process. The device first obtains an approximate location from cellular base stations, sends it to an A‑GPS server, which returns the relevant satellite parameters (visible satellites, frequencies, positions, elevation angles, etc.). The device then acquires satellites much faster, achieving a fix in a few seconds.

1.5 Ground‑Based Augmentation System

BeiDou ground‑based augmentation consists of a reference‑station network, data‑processing system, operation service platform, data broadcast system, and user terminals. Reference stations receive satellite navigation signals, process them, and broadcast the information via satellite, radio, or mobile networks to user terminals for precise positioning.

Qianxun is a global leader in precise location services, offering dynamic centimeter‑level and static millimeter‑level positioning, a key infrastructure for the IoT era.

The company was founded in August 2015 by China North Industries Group and Alibaba Group. Qianxun’s positioning is based on the BeiDou satellite system (compatible with GPS, GLONASS, Galileo) and leverages over 2,400 nationwide ground‑based augmentation stations and proprietary algorithms to provide high‑precision positioning and extended services.

2. Cellular Positioning

2.1 Overview

A base station is the radio interface through which mobile devices access the Internet; it connects the mobile switching center with the handset.

After inserting a SIM card, the device actively searches surrounding base‑station signals and establishes a connection. Multiple base stations may be visible; the strongest/nearest one is typically used for communication.

2.2 Cellular Positioning Principle

Cellular positioning principle:

1) Assume the IDs and locations of base stations are known.

When a phone detects a particular base‑station ID and signal strength, it can be estimated to be near that base station (single‑station positioning).

When a phone simultaneously detects signals from at least three base stations, triangulation can estimate its location (multi‑station positioning). Note: Triangulation treats each base station as a circle center; the radius is derived from signal strength. The intersection point gives the phone’s position.

Cellular positioning signals are prone to interference, resulting in low accuracy (hundreds of meters to 1–2 km). It cannot support turn‑by‑turn navigation.

Regardless of indoor or outdoor, positioning is fast as long as a signal exists. It is useful when GPS/Wi‑Fi are unavailable, providing a rough location.

2.3 Base‑Station Location Estimation

In section 2.2 we assumed known base‑station IDs and positions.

Base‑station modeling principle:

1) Each base‑station ID is unique and represents a specific physical device covering a limited area.

Massive terminal devices, while obtaining satellite positions, also record the connected base‑station IDs and signal metrics – this is called sampling.

All sampling points associated with the same base‑station ID are used to build a statistical model (e.g., Gaussian), thereby estimating the true base‑station location.

2.4 Cellular Positioning Error

From the modeling perspective, single‑station positioning essentially assumes the device is at the base‑station location. Since base‑station coverage ranges from hundreds of meters to kilometers, the error is large.

With the deployment of new 5G base stations, coverage areas shrink, reducing single‑station positioning error.

Note: I have experienced Huawei’s 5G small cells in large malls; they not only provide excellent communication quality but also enable meter‑level indoor positioning.

2.5 Mobile Base Stations

Mobile base‑station vehicles are deployed at large events or places needing signal reinforcement, and their movement can be observed on maps via collected positioning points.

2.6 Fake Base Stations

Devices masquerading as base stations force phones to connect to them, preventing access to public telecom networks and potentially enabling scams such as fraudulent SMS.

3. Wi‑Fi Positioning

3.1 Overview

Wi‑Fi is ubiquitous. Using Wi‑Fi signals for positioning requires no dedicated infrastructure, offering easy scalability, automatic data updates, and low cost. It was the first technology to achieve large‑scale deployment. Accuracy ranges from tens of meters to a few meters, suitable for people or vehicle navigation in hospitals, theme parks, factories, warehouses, malls, etc.

Note: Wherever Wi‑Fi is deployed, indoor or outdoor, Wi‑Fi positioning can be performed.

3.2 Wi‑Fi Positioning Principle

Wi‑Fi positioning principle:

1) Each Access Point (AP) has a unique MAC address and is assumed static over a short period.

2) Assume the positions of APs are known.

The phone collects Wi‑Fi fingerprints (e.g., AP_1,RSSI_1; AP_2,RSSI_2; …) and sends them to a Wi‑Fi positioning server.

The server applies fingerprint algorithms or AP‑based triangulation to estimate the phone’s location.

3.3 Wi‑Fi Modeling

A) Wi‑Fi data collection:

1) While walking along a map (e.g., indoor map), the phone simultaneously records Wi‑Fi information.

2) The collected Wi‑Fi fingerprints are reverse‑engineered to associate each location with its Wi‑Fi signature, establishing a mapping between fingerprint and (X, Y) coordinates.

B) Wi‑Fi modeling principle:

1) An AP’s MAC address uniquely identifies the physical device and defines its coverage area.

Massive terminals, while obtaining satellite positions, also scan Wi‑Fi fingerprints – this is sampling.

All sampling points for the same AP are used to build a statistical model (e.g., Gaussian) linking RSSI to location.

Some indoor APs cannot be scanned simultaneously with GPS; their fingerprints are correlated with GPS‑tagged APs to build a secondary model.

4. Bluetooth Positioning

4.1 Beacon

A beacon is a button‑sized device that continuously broadcasts Bluetooth signals with a unique ID. Its battery lasts 1–3 years, and retail cost varies.

Apple recommends iBeacon as the indoor‑positioning solution for iOS. Many domestic manufacturers also produce beacons.

When deployed at known locations, a device that scans a beacon’s ID and signal strength can determine its position using RSSI‑based methods or fingerprinting via a backend server.

4.2 Network‑Side Positioning

Network‑side positioning system principle:

Deploy beacons and Bluetooth gateways in the area.

The terminal enters a beacon’s range, receives the broadcast, and measures the beacon’s RSSI.

The RSSI data is sent via the gateway to a positioning server, which uses fingerprinting or other algorithms to estimate the terminal’s location.

Applications: personnel tracking, asset locating, foot‑traffic analysis, etc.

4.3 Terminal‑Side Positioning

Terminal‑side positioning system principle:

Deploy Bluetooth beacons.

Beacons continuously broadcast their IDs.

The terminal detects multiple beacons, records RSSI values and fingerprints.

The terminal sends this data to a positioning server, which computes the location using appropriate algorithms.

Applications: indoor navigation, parking‑lot vehicle locating.

5. Geomagnetic Positioning

5.1 Overview

The Earth can be approximated as a magnetic dipole near the geographic poles. The geomagnetic field consists of a primary (core‑generated, relatively stable) component and a varying component (short‑term fluctuations, weaker).

Modern steel‑reinforced concrete structures locally disturb the geomagnetic field, affecting compasses. In non‑uniform magnetic environments, different paths produce distinct magnetic observation sequences.

Consequently, walking indoors while recording magnetic sensor data yields a magnetic‑field sequence fingerprint. Later, the same device can match this fingerprint to estimate its position.

5.2 Geomagnetic Positioning Principle

5.3 IndoorAtlas

Finnish indoor‑positioning company IndoorAtlas leads in geomagnetic positioning, achieving 0.1 m to 2 m accuracy.

Because magnetometer APIs are open on Android and iOS, geomagnetic positioning breaks platform barriers.

The technique requires no physical wireless transmitters, but it relies on user motion to collect magnetic fingerprints; static points have large errors.

When a user moves, the collected magnetic fingerprint is matched to a database. Baidu invested in IndoorAtlas in 2014 and announced in 2015 that its map product integrates IndoorAtlas’s geomagnetic technology together with Wi‑Fi hotspot maps and inertial navigation.

6. Inertial Navigation

This client‑side technique uses motion sensors (accelerometer, gyroscope, magnetometer, barometer) to measure velocity, direction, and acceleration. By dead‑reckoning, the device’s position is computed.

Errors accumulate over time, so external high‑precision data (e.g., GPS outdoors, Wi‑Fi or Bluetooth indoors) are needed for periodic calibration.

7. LED Positioning

LED devices encode an ID into light. The light continuously broadcasts the ID, which a phone’s front‑camera captures, uploads to a server, and decodes to obtain a position.

Because no extra infrastructure is required, scaling the number of terminals does not affect performance, and high precision can be achieved.

Visible‑light positioning is deployed in many North American and European malls; a user can locate a shelf by detecting the light signal from the shelf’s LED.

8. Vision Positioning

Vision positioning systems fall into two categories:

1) Mobile sensors (cameras) capture images to determine the sensor’s pose.

2) Fixed sensors determine the position of a target within the captured image.

Reference 3D building models and images are used to compare against captured frames.

To improve robustness, pre‑deployed visual markers (e.g., QR codes) can be placed.

Reference points can be projected onto the environment based on pre‑deployed markers.

Fusion with other sensors (e.g., IMU, Wi‑Fi) enhances accuracy, coverage, and robustness.

9. UWB Positioning

Ultra‑Wideband (UWB) is a new wireless communication technology that differs greatly from traditional carrier‑based systems.

UWB transmits and receives ultra‑narrow pulses with nanosecond or sub‑nanosecond duration, providing a bandwidth of 3.1–10.6 GHz. It offers high data rates, low transmit power, strong penetration, and operates without a carrier.

UWB positioning uses pre‑deployed anchor nodes with known locations. A blind node communicates with these anchors, and triangulation or fingerprinting yields a position with meter‑level accuracy.

10. IP Positioning

10.1 Overview

IP positioning determines location based on the IP address contained in a request.

10.2 IP Positioning Principle

Besides purchasing IP‑address databases from operators, one can build an IP database.

The IP database is constructed by using massive terminal devices that simultaneously obtain satellite positions and record their IP addresses, thereby establishing a real‑time mapping between IP and geography.

Combining IP data with device/account information can improve accuracy. Fixed‑line IPs can achieve decimeter‑level errors, while dynamic mobile IPs may have errors spanning cities.

The End

If you found this content insightful, please give a light tap:

1. Click "Reading" to let more people see this valuable technical article.

2. Follow the public account "Hello Tech Team" to receive the latest technical posts promptly.

If you like it, give a 👍. Your support fuels us to keep delivering valuable technical shares.