Location fingerprinting - what is it, and should you choose it as your IPS technology?

When choosing an indoor location provider, there are a few key elements that you need to look into: the accuracy of the blue dot, the scalability of the solution, the flexibility to adapt to an always-changing floorplan, and finally, ensuring there are real-world case studies to back the technology. Over the years, some companies have claimed they can achieve that by leveraging wireless fidelity via location fingerprinting, but can they back up their claims? 

In this article, we'll explore what location fingerprinting is and whether it lives up to its hype. We'll also explain why at Pointr, we don't use location fingerprinting and how we achieve highly accurate and scalable location technology.

You'll discover

  • Insights into location fingerprinting, including a detailed explanation of what it is and how it works
  • The drawbacks of using location fingerprinting for indoor positioning and why Pointr avoids this technology
  • Real-world examples of location fingerprinting 
  • How do we avoid the issues associated with location fingerprinting by using a completely different approach in the Pointr Deep Location® platform?

Location fingerprinting - what is it and how does it work?

Fingerprinting is a common indoor positioning technology to determine a user's position. The technique relies on signal strength data, called "RSS (Received Signal Strength)," which represents the distance of each beacon or sensor from a user's device. 

This technique involves creating a signature of the venue by walking around the venue step by step and recording signals at every step. By turning the venue into a grid, it is possible to create a database of different signal strengths to know which signal strengths to attribute to each spot within a building. 

By cross-referencing the signal strength (RSS) against the pre-existing record, the system can assume the device's "live" position by calculating the distance between a beacon/ sensor and a user device using the fingerprint data when the system was first set up. 

Location fingerprinting usually consists of two main phases - the offline training and the online testing phases.

The offline training phase trains the fingerprinting algorithm to learn the RSSI (Received Signal Strength Indicator) at various points, including "Access Points (APs)," "Reference Points (RPs"), and the "User Devices" - each of them acting as a landmark in the indoor environment.  To train the algorithm, you need to physically walk around the venue and "record" the signal strength data on a mobile device. The collected data is then stored in a database along with their location coordinates, called "Reference Points."

What happens in the database

  • Record and store wireless fingerprints data (RSS value)
  • Executes matching algorithm to match reference points 
  • Calculate the devices' location and update the live data

 

The online matching phase measures the RSSIs from beacons and compares them with the values stored in the database to infer a location. The Received Signal Strength (RSS) value obtained at any point in a location is called a location fingerprint. 

When a user moves around the indoor environment, the following process will happen:

  • An access point (AP) or beacon sends the target signal
  • A user's device detects the emitted signal
  • APs record the signal strength and upload the fingerprints of a user's device (RSS value) to the database
  • The database matches the received reference points (user's device) with the stored ones (from the offline training phase) to estimate a device's "live" location

The pros and cons of location fingerprinting for indoor positioning

Fingerprinting is probably the most common technique in use today. Companies like Apple and Google use Wi-Fi fingerprinting to provide indoor location with 15-20 meters accuracy. 

The main advantage of a fingerprinting system is that it's relatively straightforward technically. If enough beacons are calibrated correctly, then the algorithm only needs to look up multiple signal strengths and use those to triangulate a device's location. There are no complex algorithms required. 

Want to learn everything there is to know about indoor positioning? Download our guide.

Get the Guide

 

However, fingerprinting has numerous drawbacks compared to a complex location platform such as Pointr Deep Location®, which makes complex and real-time calculations as devices or users move throughout an indoor environment. 

The main disadvantages include:

  • Fingerprinting is not suitable for venues that update their layout - to use location fingerprinting, you need to record the venue to collect the reference points physically. Suppose you own a building that require regular layout changes, e.g., hybrid workplaces, busy hospitals, retail floors, and airports. You would need to go through the recording process and re-calibrate the system each time you change the layout to ensure accuracy, making location fingerprinting impractical for venues that change regularly.
  • The technique only works online - that’s because the signature is kept on a server. This means if a visitor doesn’t have data or if there’s no signal in the user’s location, the system will have to wait until they are back online before they can load their location and provide wayfinding capabilities. This is an issue for areas with limited internet access where there is no signal, or at international airports, where visitors typically don’t have data.

Beacon Position and Signal - Bad ExampleHD

  • Cumbersome to install and maintain  - location fingerprinting is difficult to set up. It requires walking step by step to configure, and if there’s any mistake in the calibration, the entire process needs to be started again. It is also difficult to maintain as it needs to be recalibrated every time there is a small layout change, for instance when you move a metal cabinet or close an entrance. 
  • Lack of accuracy and stability - the performance of fingerprint-based methods depends on the number of beacons or sensors (APs) to populate accurate signal strength readings (RSSI). As its performance relies on measuring the signal strength, some factors might affect its accuracies: radio interference caused by building material, layouts or object mobility (inc. human bodies), device orientation, RSSI consistency, and more APs deployed, it can cause overlapping radio channel issues.

Material

Decline in RSS power level (dB)

Plasterboard

3-5

Glass wall & metal frame

6

Metal tool

6-10

Window

3

Concrete wall

6-15

Block wall

4-6

Radio interference caused by building material. Source: Machine Learning-Based Indoor Localization Using Wi-Fi RSSI Fingerprints: An Overview

Radio interference caused by moving objects (e.g. human bodies). Source: Swiss Federal Institute of Technology.

  • Cost - In order to improve the accuracy and stability of the positioning, fingerprinting would rely on increasing the number of APs. This can substantially increase positioning costs. On top of that, every time the venue layout changes, a complete recording and training process must be repeated to ensure accuracy. This would likely generate more cost as well.
  • Not easily scalable - location fingerprinting doesn't scale well for all of the reasons mentioned above - it would be a very difficult task to record signatures of dozens, hundreds, or even thousands of buildings in order to digitize them with indoor location services. In fact, there isn’t a single example of fingerprinting deployment at scale.Challenges of fingerprinting-based indoor localization.

What are the real-world examples of location fingerprinting?

Companies have tried leveraging location fingerprinting for indoor location services and claim they can achieve a high-performance indoor positioning system. Let's take a look at some of the real-world examples:

Ekahau

As one of the largest Wi-Fi/WLAN Network suppliers globally, Ekahau has developed indoor location technology using location fingerprinting. They've been making very bold claims about accuracy and performance, and it seems that they are one of the market leaders for indoor tracking systems. However, the most significant pain points for location fingerprinting still remain - flexibility and scalability. When reconstruction happens, the users need to re-calibrate both the APs and the data, including conducting a full site survey and recording process to ensure the accuracy and consistency of its positioning systems. 

Skyhook wireless

One of the pioneers in location positioning technology, Skyhook wireless has developed its indoor location technology using location fingerprinting with a mix of infrastructure, including Wi-Fi, GPS, and device sensors. Their solution claims to deliver 5-8 meters of accuracy but recommends having as many APs as referenced to ensure accuracy.

To solve the biggest challenge for location fingerprinting, they developed a self-healing database system to tackle scalability and flexibility issues, allowing the system to identify if a new AP is added or moved and reconfigure the positions of APs directly in the backend without the need to conduct physical recordings or calibrations on-site. However, spoofing APs increases the risk of jamming other APs and breaking down the system.

No location fingerprinting here at Pointr, 100% guaranteed!

Pointr doesn't use location fingerprinting - essentially, fingerprinting is challenging to set up and maintain, inaccurate, expensive, and incapable of scaling quickly and easily.

Instead, the Pointr approach to indoor positioning is to use machine learning algorithms to work out where the blue-dot should be and the correct path and orientation to lead the user (pathfinding), using calculations relating to signals emitted by beacons, sensors, Wi-Fi access points, and smart lighting.

Pointr's Deep Location® solution is smart enough to calculate the real-time location of a user directly on their device via Pointr SDK, which means it doesn't need to do any location fingerprinting. Pointr's solution can even work offline without any internet access, making it the best solution for areas like parking lots, basements, or places with limited internet access.

Pointr Deep Location® platform is also hardware agnostic, and it works with various wireless devices with Bluetooth function, including beacons, sensors, Wi-Fi AP, or intelligent lighting. It's is easy to deploy with the minimum maintenance requirement.

Our approach ensures the Pointr Deep Location® platform is cost-effective, accurate, and flexible enough to scale quickly while protecting your buildings from the risks associated with location fingerprinting.

 

Deep Location platform

Location fingerprinting

Location accuracy

< 1-3 meters, real-time

5-8 meters, delayed 

(depend on signal strengths)

Scalability

Machine-learning based algorithm adapt to layout changes 100%

Require on-site survey and data collection every time layout change

Consistency

Work in offline mode

Need internet access

Hardware requirement

Hardware agnostic

Only work with WiFi access points

Time to deploy

Hours

Weeks or months

Upkeep

None

Monthly maintenance

Setup cost

$

$$$

 

Deep Location® is Pointr's ground-breaking take on indoor location and indoor navigation, which revolves around being scalable and software-focused. Our product can be deployed quickly in an indoor environment when compared to other solutions that don't have that depth of ability.

 

Find out more in this 8-minute clip from the Mr. Beacon podcast

Watch the interview

 

About Pointr

Pointr is a global leader in indoor location. Pointr's Deep Location® software technology uses machine-learning techniques to create the best performing and the only scalable indoor location technology available today. Deep Location® enables location-based services such as digital mapping and smooth indoor-outdoor wayfinding. 

We work with major customers in healthcare, smart workplace, retail, and aviation across North America, Europe, and Asia, including UCHealth, international corporations (CBRE), the U.S. Department of Homeland Security, U.S. Airports (Washington Regan and National), two major U.S. Airlines and one of the major U.S. department store retailers across 2,000+ locations with millions of mobile application users.

Related topics:

Author:

les Les Blythe
eva Eva Cheng