Category: Location

Making Sense of Indoor Location

There’s a recent research paper on Indexing for Moving Objects in Multi-Floor Indoor Spaces That Supports Complex Semantic Queries. It says humans spend 87% of their time in indoor spaces such as private residences and office buildings and it’s becoming more important to be able to derive meaning from indoor location.

The paper explains how outdoor moving object management technology, which is very mature, cannot be applied to indoor spaces. Instead you need software that not only understands floors but also multi-floors and inter-floor (elevators and stairs) cells. The paper describes an index that can store indoor moving objects in multi-floor indoor spaces that can support 3D spatial queries.

Beacon Locating Accuracy

There’s a useful article by Steffen von Bünau of Kontakt on Real Time Location Systems (RTLS). Steffen says:

“Accuracy is an expensive vanity metric unless it is necessary to get the job done.”

Most scenarios don’t usually need very accurate positioning and creating unnecessarily accurate systems is expensive. Steffen doesn’t say why they are expensive but one of the article’s comments provides an answer. Ultra wideband based RTLS is expensive compared to Bluetooth LE.

Also, accurate systems tend to need calibration that’s time consuming and costly in human resource. Calibration implies tuning to a particular physical and wireless environment. If the environment changes then so might the calibration.

The required accuracy of a RTLS should be derived from the business requirements.

Improving on Beacon Immediate, Near and Far

We recently highlighted an article on Beacon Trajectory Smoothing. Faheem Zafari, Ioannis Papapanagiotou, Michael Devetsikiotis and Thomas Hacker have a new paper on An iBeacon based Proximity and Indoor Localization System (pdf) that also uses filtering.

They use a Server-Side Running Average (SRA) and Server-Side Kalman Filter (SKF) to improve the proximity detection accuracy compared to Apple’s immediate, near and far indicators.

The researchers found:

The current (Apple) approach achieved a proximity detection accuracy of 65.83% and 67.5% in environment 1 and environment 2 respectively. SRA achieved 92.5% and 96.6% proximity detection accuracy which is 26.7% and 29.1% improvement over the current approach in environment 1 and 2 respectively

What’s interesting here is that the researchers have quantified the accuracy of Apple’s implementation in two scenarios. The accuracy isn’t that good and as the researchers have shown, can be improved upon significantly.

Beacon RTLS Accuracy

Steffen von Bünau of Kontakt.io has an interesting article on “Jobs to be done – Accuracy in Real Time Location Systems”. He asks:

Who will use the information and what is to be achieved with it?

He questions whether organisations need room level accuracy or location within a room.

As Steffan says, trilateration can be used for positioning within a room.

However, determining accurate location within a room is much harder and more expensive to achieve and needs fingerprinting. Fingerprinting involves going over the target area to sample beacon signal strengths that’s time consuming. It’s also the case that the more you tune these things, the easier it is that they can go out of tune when the environment changes. New or changed items in a room can easily change signal strength readings and cause the need for re-fingerprinting.

As Steffen says:

“Accuracy is an expensive vanity metric unless it is necessary to get the job done.”

Read more about beacons for RTLS and our BeaconRTLS platform.

Read about BluetoothLocationEngine™

Beacon Trajectory Smoothing

The problem with using RSSI for detecting location is that raw data contains lots of noise. Also, this noise becomes more prevalent in the viewed data when location samples are taken less often. There’s a useful new article at InfoQ on Processing Streaming Human Trajectories with WSO2 CEP.

The idea uses Kalman filtering to smooth noisy human trajectories.

Before and after filtering

This method is particularly useful for large realtime IoT rollouts because it uses a very small memory resource, is very fast and the calculation is recursive, so new values can be processed as they arrive.

Beacon Location Accuracy

There’s some recent new research on ‘Analysis of Object Location Accuracy for iBeacon Technology based on the RSSI Path Loss Model and Fingerprint Map’ by Damian Grzechca, Piotr Pelczar, Łukasz Chruszczyk.

They evaluated RSSI and indoor positioning trilateration algorithms in order to determine location accuracy. After lots of experimentation and mathematics, they calculated the average error to be 1.09m for 1–9m and 1.75m for 1-20m and after trilateration an average error 2.45m was achieved.

The conclusions give some hints how better results might be achieved. For example, correlating the RSSI with accelerometer, gyroscope and other sensors. Other strategies might be to excluding areas where an object
cannot move, or filtering out situations where objects move but accelerometer measurements don’t match.

Crowd Analysis Using Beacons

With so many uses of beacons centred around notifications to users, it’s interesting to see Queen Mary University of London doing something different. Research by Kleomenis Katevas, Laurissa Tokarchuk, Hamed Haddadi and Richard G. Clegg of the Department of Computing of Imperial College looks into detecting group (crowd) formations using iBeacon (pdf).

They used beacon RSSI and phone motion together with algorithms based on graph theory to predict interactions inside the crowd. They verified their finding using using video footage as ground truth.

distanceestimationmodels

The paper has some particularly interesting observations from testing RSSI in an EMC screened anechoic chamber and also has some information on distance estimation models.