Using AI Machine Learning with Bluetooth Angle of Arrival (AoA)

There’s new research from Universities in Piraeus, Greece and Berlin, Germany, together with U-Blox AG in Switzerland who create Bluetooth Angle of Arrival prototyping boards on Deep Learning-Based Indoor Localization Using Multi-View BLE Signal.

Processing of Bluetooth Angle of Arrival usually requires radiogoniometry spectral analysis of radio in-phase and quadrature-phase (IQ) signals in order to then determine location by triangulation. Instead, this paper proposes machine learning of IQ and signal strength (RSSI) data from multiple anchor points to determine location. AoA processing also uses distributed processing across the anchors to improve performance.

The developed machine learning models were found to be robust against modifications of room furniture configurations and materials and it’s therefore expected that they have high re-usability (machine learning generalisation) potential. The system achieved a localization accuracy of 70cm.

Read about BluetoothLocationEngine™

Read about PrecisionRTLS™

Bluetooth for Locating

The Bluetooth SIG, the organisation that produces Bluetooth standards, has a recent post The Myths & Facts About Bluetooth Technology as a Positioning Radio. It talks about the location services in general and how they have evolved over time. It explains how Bluetooth helps solve key enterprise pain points to save tens to hundreds of billions of dollars globally through enhanced operational efficiencies, increased worker safety, and loss prevention.

In manufacturing facilities, billions of dollars are lost through unplanned downtime thanks to being unable to locate assets, tools, and equipment. In warehouses, RTLS can help automate the tracking of assets, such as pallets, which is becoming more essential with the ever-increasing size, complexity, and amount of assets stored

Despite the gains thus far, this only represents as small proportion of the opportunity because only a very small percentage of the potential addressable market in the enterprise is using RTLS.

The article continues with a summary of the myths we covered in a previous post.

ABI Research expects that will be a 2.5x increase in total Bluetooth RTLS deployments over the next five years, with the fastest growing segments being healthcare, warehouse and logistics, manufacturing and smart building.

Finding the Nearest Beacon

There’s new research from Faculty of Electrical Engineering and Computer Science, University of Maribor, Slovenia on Improved Bluetooth Low Energy Sensor Detection for Indoor Localization Services.

While there has been lots of research into server-side processing to improve location accuracy, this research instead looks into improving accuracy locally, in terms of finding the nearest beacon. This kind of processing is often needed where smartphone apps provide users with contextual information based on their location, for example, in museums.

It’s not possible to use the raw received signal strength (RSSI) because it changes frequently due to changes in blocking and reflection in a room. Any errors in determining the correct transmitter can cause errors in displaying relevant information which, in turn, leads to a poor visitor user experience.

The study involved use of iBeacons detected by Android smartphones, both in a controlled room with three obstacles and a real-world setting Expo Museum.

The proposed algorithm stabilised the RSSI by considering previous measurements to filter out sudden fluctuation of the RSSI signal or the rapid movement of the mobile device. The smartphone’s accelerometer was also used dynamically change the scan interval based on the user’s movement.

In the controlled room, the proposed algorithm had a 14.29% better success rate than a standard algorithm using the raw RSSI values. It performed particularly (20%) better in spaces having medium or high density of physical obstacles. It also performed better in the real-world Expo environment with a success rate of 95% compared to 87% with a standard algorithm.

Bluetooth Myths and Facts

There’s a useful new webinar at the Bluetooth SIG on The Myths & Facts About Bluetooth® Technology as a Positioning Radio. Fabio Belloni from Quuppa explains the main Bluetooth myths and facts:

  • Performance – There are misconceptions about accuracy, latency and reliability brought over from older systems using only received signal strength (RSSI). Newer systems based on Bluetooth direction finding provide much improved performance.
  • Communication Range & Coverage Area – People incorrectly think Bluetooth is a short range 10m – 15m technology. This isn’t so. Long range beacons can transmit up to 1.5Km and can work up to 100m in location finding scenarios.
  • Multipath Propagation – It’s wrongly perceived that Bluetooth is poor in harsh environments. Bluetooth is, in fact, designed for factory floor and additionally newer AoA direction finding can use spectral analysis to reduce the affect of radio reflections.

Gabriel Desjardins from Broadcom mentions how location technologies have overcome the peak of inflated expectations caused by UWB and are now in the plateau of productivity provided by Bluetooth LE.

Andrew Zignani shows the results of a survey on RTLS from 213 C-Level decision makers across five main verticals. Only 13% of businesses have already deployed RTLS and there will be a increased uptake over the next 5 years. Technology fragmentation and operational/maintenance cost are incorrectly seen as the barriers to adoption. The new Bluetooth AoA direction finding standard is easing fragmentation. The maintenance cost is actually very low compared to the ROI in most scenarios. Most want beacon battery life to be 90+ days and cost to be $11-$20 that are easily achievable with today’s beacons.

Asset Tracking For Manufacturers

Today’s just-in-time and busy manufacturing processes means that manual tracking of pallets for inbound and outbound shipments often can’t keep pace with the speed of production. Production and assembly requires the quick locating of components. Delays and inaccuracies due to lost components lead to increased costs, employee frustration and ultimately customer disappointment.

Competitive pressures are also driving the need to reduce labour thus reducing the capacity to manually search for items. Customisation using configured options and demand-driven production is also increasing the degree of inbound component searching that exacerbates the problems.

Even those companies using legacy tracking solutions find that location is only as good as the last barcode or RFID scan. Humans get lazy, make mistakes and don’t scan, causing pallets, crates and boxes to get lost. Many RFID readers don’t work reliably near metal components. Relying on a system that can’t find just a few items can be worse that a manual system that works but is slower. Bluetooth asset tracking solves these problems because the location is automatically collected in real-time and is continually updated.

Asset tracking can be applied to items such as components, pallets, cases, tools, returnable assets such as racks and cages as well as items on loan to ensure they are returned on time. It can improve worker safety and provide alerts in cases of congestion, perimeter deviation and lone worker distress. It can ensure forklifts are being fully utilised, are taking an optimum route, haven’t crashed into racking and haven’t gone out of an area.

The real-time visibility allows connected systems to generate confirmation and exception alerts and automatically trigger shipping processes, replacing costly manual workflows. Tracking outputs also allows confirmation that the correct things are loaded on the correct transport.

A Bluetooth-based real time location system (RTLS) increases visibility and allows the manufacturing process to adapt in real-time to short term business needs. It provides cost savings, greater efficiency and business intelligence that can be used to derive larger scale changes based on data rather than gut instinct. Overall reporting of input and outputs provides input to management reporting to monitor the business.

Read about BeaconRTLS™
Read about PrecisionRTLS™

2D vs 3D Bluetooth AoA Direction Finding

Current AoA locators only have antennas in one plane which means they can only provide angles in two (elevation and azimuth) dimensions. A locator therefore sees assets as being somewhere along an imaginary line or ray emanating from the locator.

If the height, perhaps of a worn lanyard, is known and tends to not change much then it’s possible to estimate the 3D location. Obviously, if the person climbs some steps for kneels down then the location becomes less accurate in all dimensions.

The other solution is to use multiple locators to triangulate two or more locator lines. This is more accurate because it doesn’t rely on a known average height and provides the opportunity to use more than two locators to increase accuracy still further.

3D provides the best accuracy. 2D location allows use of fewer locators with the trade off of less accuracy. For example, the four locators in the Minew AoA kit can be placed in different rooms or areas rather than covering an overlapping area. 2D location also has the implicit advantage of supporting more beacons because the locators and subsequent systems are doing less work.

Read about BluetoothLocationEngine™
Read about PrecisionRTLS™

Bluetooth Direction Finding Antenna Arrays

Bluetooth direction finding Angle of Arrival (AoA) uses multiple antenna in one device that uses the phase difference of signals received at different antenna to determine the angle and hence location of a beacon.

We are seeing a variety of designs but most use printed circuit board patches for antennas for reasons of cost and compactness.

All these designs use a radio frequency switch that switches each antenna, in turn, to just one Bluetooth chip to save cost and complexity. You can see this in some of the designs as tracks leading from each antenna to one chip and then one track from that chip to the Bluetooth system on a chip (SoC). The switch is very fast, of the order of 1 microsecond, to capture the same origin signal across all antennas.

Take care to purchase production-ready hardware. While there are currently many antenna array designs, some are just prototype or reference boards not intended for production. The software accompanying prototype or reference boards is also tends to be non-existent and in cases where it does exist, it won’t scale to more than a few beacons.

In practice, a location engine employing AoA radiogoniometry is required to process the radio signals from the Bluetooth SoC. The radio signals are also wirelessly noisy and have to be processed to mitigate reflections, interference and signal spread delays. Additional processing is needed to triangulate the angles from multiple locators. All this isn’t trivial given that the algorithms are computationally expensive and have to be executed extremely quickly to support a large number of beacons.

Minew AoA Kit

Read about BluetoothLocationEngine™
Read about PrecisionRTLS™

RSSI vs AoA Bluetooth Asset Tracking

In a previous post on iBeacon Microlocation Accuracy we explained how assets can be tracked using Bluetooth Received Signal Strength (RSSI) or Angle of Arrival (AoA). We advised working out what accuracy is needed prior to seeking out an appropriate solution. However, accuracy isn’t always the only consideration and here is a more complete list of factors.


RSSI asset tracking can achieve accuracies of about 1.5m within a shorter range confined space and 5m at the longer distances. RSSI zone-based systems where beacons are found to the nearest gateway, are accurate to the inter-gateway distance that can be of the order of cm. However having such as large gateway density is usually only practical for very small areas.

AoA asset tracking achieves sub-metre accuracy. The accuracy depends most on the distance between the locator and beacon but is also affected by the locator hardware quality, radio signal noise, surfaces causing radio reflections, the accuracy of locator placement and beacon orientation.

Maximum number of beacons

AoA-based asset tracking produces and requires much more data which means the locators and software systems have to deal with more data. The data throughput for both types of system depends on the required minimum latency that in-turn depends on how often the beacons advertise. RSSI-based systems support up to high tens of thousands of beacons while AoA supports thousands of beacons.

Beacon variety and IoT

RSSI-based systems can use any beacon and hence support a large range of sensor beacons that can detect movement (accelerometer), movement (started/stopped moving), button press, temperature, humidity, air pressure, light level, open/closed (magnetic hall effect), proximity (PIR), proximity (cm range), fall detection, smoke, natural gas and water leak.

AoA beacons are more specialised and currently only support limited IoT sensing such as movement (accelerometer) and button press.


AoA locators, gateways and beacons are more complex and are therefore more costly. AoA also needs more locators/gateways per sq area. Hence, AoA systems are x3 to x4 more expensive than RSSI systems.

Setup effort

The accuracy of AoA requires that locators be more carefully positioned than for RSSI, in particular the site and AoA locator positions need to be carefully measured.

Beaconzone supplies both RSSI and AoA systems. Contact us to determine the best type of system for your needs.

Using Bluetooth Mesh and Space, Time, Frequency Diversity to Improve Locating Accuracy

There’s new research from the Department of Electrical Engineering, University of North Texas, USA on Measurement and Analysis of RSS Using Bluetooth Mesh Network for Localization Applications. The paper explains how received signal strength (RSSI) based solutions have accuracy limitations in radio multipath (radio reflective) environments. It describes a solution that improves accuracy using Bluetooth mesh and Bluetooth channel-based processing.

Bluetooth and WiFi Channels

A system was created that exploits the space, time, and frequency diversities in measurements. Different Bluetooth channels have different fading effects.

Advertising was modified to make it Bluetooth channel-aware to be able to differentiate the fading effects. It was possible to reduce the residual fitting errors in the path loss models by using a space-time-frequency diversity combining scheme.

The system was demonstrated using ESP32 BLE modules.

The system significantly reduced the residual linear regression fitting errors in path loss models. It was able to more accurately use RSSI to measure the distance between the transmitter and receiver. The researchers demonstrated it’s possible to implement the proposed multi-receiver configuration and the diversity combining scheme using commercial off-the-shelf standard BLE devices.

Probabilistic vs Neural Network iBeacon Positioning

There’s new research by ITMO University, Russia on the Implementation of Indoor Positioning Methods: Virtual Hospital Case. The paper describes how positioning can be used to discover typical pathways, queues and bottlenecks in healthcare scenarios. The researchers implemented and compared two ways to mitigate noise in Bluetooth beacon RSSI data.

The probabilistic and neural network methods both use past recorded data to compare with new data. This is known as fingerprinting. The neural network method is less complex when there’s need to scale to locating many objects. The researchers tested the methods at the outpatient department of the cardio medical unit of Almazov National Medical Research Centre.

Comparison of the methods showed they give approximately the same error of between 0.96m and 2.11m. However, the neural network-based approach significantly increased performance.