Movement Constraint-based Location Tracking

Researchers at the Pusan National University, Korea have a new paper on Applying Movement Constraints to BLE RSSI-Based Indoor Positioning for Extracting Valid Semantic Trajectories.

The received signal strength (RSSI) of beacons is often used to infer location. However, the RSSI is subject to reflection and blocking from walls, people and other obstacles causing the derived locations from the raw data to be ‘jumpy’. There are many ways to process the raw data, such as Hidden Markov Models, k-nearest neighbors and Deep Neural Networks (DNN) to obtain smoother trajectories.

The researchers use movement constraints and sliding-window aggregation to extract invalid trajectories and provide real-time semantic trajectories.

The paper shows the proposed movement constraint-based approach extracts valid trajectories that are comparable to the unconstrained and non-machine language approaches. This new approach is particularly suited to dynamic indoor environments where the reflection and blocking changes over time.

Comparison of Bluetooth LE Locating Methods

There’s research just published on A Comparison Analysis of BLE-Based Algorithms for Localization in Industrial Environments. The research compares trilateration, fingerprinting and a machine-learning based k-nearest neighbors regressor for determining the location from signals from multiple beacons.

Multi-layer perceptron (MLP) schematic model
Error box plots for the three fingerprinting algorithms with different beacon densities. Results for a fingerprint grid with one measurement every 0.5 m.

The results show fingerprinting is better than distance-based schemes in industrial environments due to the presence of large moving metal objects that shadow and reflect wireless signals. The three methods were found to provide similar localisation accuracy. The authors say the machine learning method is best due to less complexity and better adaptability. The machine learning method does not need regular calibration as is the case with fingerprinting.

Read about Determining Location Using Bluetooth Beacons

Using Bluetooth and WiFi RSSI for Locating

There’s a recent paper by Hongji Cao,Yunjia Wang,Jingxue Bi and Hongxia Qi of China University of Mining and Technology on An Adaptive Bluetooth/Wi-Fi Fingerprint Positioning Method based on Gaussian Process Regression and Relative Distance.

The paper looks into how to combine both Bluetooth fingerprint positioning (BFP) and Wi-Fi fingerprint positioning (WFP) to provide for an adaptive Bluetooth/Wi-Fi fingerprint positioning system based on Gaussian process regression (GPR).

The adapative feature is particularly useful because fingerprint acquisition requires a great deal of effort and requires subsequent update and maintenance.This new method provides a better positioning than Bluetooth and Wi-Fi positioning alone but at the cost of extra computation.

Read about Using Beacons, iBeacons for Real-time Locating Systems (RTLS)

Using Beacons for Race Timing

There’s novel recent research on City Marathon Active Timing System Using Bluetooth Low Energy Technology by Chun-I Sun, Jung-Tang Huang, Shih-Chi Weng and Meng-Fan Chien of Taiwan.

The authors discuss the use of beacons vs RFID and create a system using Received Signal Strength Indicator RSSI and gateways connected to detector mats:

Beacons are carried by athletes. The gateways sync their times via NTP and send data up to a MongoDB database:

An accuracy of ±156 ms was achieved which compares well to the nearest second used to generally record times and resolution accuracy of 0.1s for commercial transponder timing systems.

The Affect of Transmission Power, Advertising Interval and Beacon Placement Density on Location Accuracy

There’s recent research by Gabriele Salvatore de Blasi, José Carlos Rodríguez-Rodríguez, Carmelo R. García and Alexis Quesada-Arencibia of University of Las Palmas de Gran Canaria, Spain on Beacon-Related Parameters of Bluetooth Low Energy: Development of a Semi-Automatic System to Study Their Impact on Indoor Positioning Systems.

The paper starts by giving an overview of fingerprinting. It explains how fingerprinting is time-consuming and labour-intensive. Fingerprinting is affected by:

“Reflection, refraction, path loss, large fluctuations, multipath fading, non-line-of-sight (NLOS) conditions”

A problem is that some environment-related factors change over time, such as changes in hardware/furniture, the presence of people and ambient humidity conditions meaning that fingerprinting isn’t a one-off activity.

The researchers conclude that the highest transmission power (+4 dBm) produces the best location accuracy. However, this uses a lot of battery power. Use of the lowest power (−20 dBm) only worsened the accuracy by 11.8%. Similarly, lowering the density of the beacons by around 50%, the error increase was only about 9.2%. Increasing the advertising interval didn’t have a significant impact on the accuracy.

The affect of beacon orientation was assessed and vertical orientation was found to be best. Read our previous article on orientation. As expected, introducing a slight electrical noise during the positioning phase did not significantly affect accuracy.

Read about Locating with Beacons

Using iBeacons with Intelligent Displaying and Alerting Systems

There’s recent research into using iBeacons with intelligent displaying and alerting systems (SICIAD) typically found in public buildings and offices. The paper An Intelligent Low-Power Displaying System with Integrated Emergency Alerting Capability by Marius Vochin, Alexandru Vulpe, Laurentiu Boicescu, Serban Georgica Obreja and George Suciu of the University of Bucharest shows how beacons can be used to determine indoor position of mobile terminals or signalling points of interest.

An Android app uses the beacons to detect location and sends it to the SICIAD system. The researchers concluded that:

“By using an appropriate number of beacons and optimal positions, a relatively precise indoor localization can be obtained with iBeacon technology”

Beacon Based Navigation for the Visually Impaired

There’s a useful recent research paper by Basem AL-Madani, Farid Orujov, Rytis Maskeliūnas, Robertas Damaševičius,and Algimantas Venčkauskas on Fuzzy Logic Type-2 Based Wireless Indoor Localization System for Navigation of Visually Impaired People in Buildings.

The paper explores indoor location algorithms and implements a fingerprinting system using RSSI that achieves an average error of 0.43m.

The authors’ ‘fuzzy logic type-2’ system allows for complex environments such as buildings with glass/metal corridors. They comment that fingerprinting requires pre-configuration which is one of the main disadvantages of this method.

The Affect of Power Levels on Wireless Indoor Localisation Accuracy

There’s new research by Umair Mujtaba Qureshi, Zuneera Umair and Gerhard Petrus Hancke of the Department of Computer Science, City University of Hong Kong on Evaluating the Implications of Varying Bluetooth Low Energy (BLE) Transmission Power Levels on Wireless Indoor Localization Accuracy and Precision. The paper takes a deep look into the relationship between transmitted power and signal stability. It also looks at ways of filtering received signal strength (RSSI) data to improve the location accuracy.

The main insight is that along with the expected difference in the RSSI attenuation there is a considerable difference in the BLE signal variation at all transmission power levels with respect to distance. The variation increases and the localisation accuracy decreases from high to low transmission power levels:

Another observation is that outliers in the data tend to affect the localisation accuracy. Applying filters to the data, they achieved a location accuracy of 2.2 meters with a precision of 95%.

One comment we have is that the researchers didn’t try different beacons. As we mentioned in 2016, the RSSI stability also varies across different beacon models.

Need more help? Consider a Feasibility Study.

Bluetooth Positioning Using Separate Bluetooth Channels

While we wait for commercial Bluetooth 5.1 direction finding solutions to become available, people are trying to refine traditional locating methods to gain more accuracy. Baichuan Huang, Jingbin Liu, Wei Sun and Fan Yang have a research paper on A Robust Indoor Positioning Method based on Bluetooth Low Energy with Separate Channel Information.

They have observed that the stability of the received Bluetooth signal strength RSSI depends on which Channel 37, 38 or 39 the signal is being received on. This is because the channels slightly overlap the WiFi channels and there can be other Bluetooth devices also using the same channels.

The method analyses the channels over time and chooses those it thinks has least interference and most stable RSSI. This reduces the positioning error by 0.2m, to 2.2m, at a distance of 3.6m.

Read about Determining Location Using Bluetooth Beacons

Using AI Machine Learning on Bluetooth RSSI to Obtain Location

In our previous post on iBeacon Microlocation Accuracy we explained how distance can be inferred from the received signal strength indicator (RSSI). We also explained how techniques such as trilateration, calibration and angle of arrival (AoA) can be used to improve location accuracy.

There’s new research presented at The 17th Annual International Conference on Mobile Systems, Applications, and Services (MobiSys ’19) by researchers from Nagoya University, Japan that looks into the use of AI machine learning to process Bluetooth RSSI to obtain location.

Their study was based on a large-scale exhibition where they placed scanning devices:

They implemented a LSTM neural network and experimented with the number of layers:

They obtained best results with the simplest machine learning model with only 1 LSTM:

As is often the case with machine learning, more complex models over-learn on the training data such that they don’t work with new, subsequent data. Simple models are more generic and work not just with the training data but with new scenarios.

The researchers managed to achieve an accuracy of 2.44m at 75 percentile – whatever that means – we guess in 75% of the cases. 2.44m is ok and compares well to accuracies of about 1.5m within a shorter range confined space and 5m at the longer distances achieved using conventional methods. As with all machine learning, further parameter tuning usually improves the accuracy further but can take along time and effort. It’s our experience that using other types of RNN in conjunction with LSTM can also improve accuracy.

If you want to view the research paper you need to download all the papers from the conference (zip) and extract p558-uranoA.pdf. Some of the other papers also make interesting, if not directly relevant, reading.

Read about AI Machine Learning with Beacons