How Bluetooth direction finding can unlock precision indoor positioning - Embedded.com

How Bluetooth direction finding can unlock precision indoor positioning

Version 5.1 of the Bluetooth specification introduced ‘direction finding’. This represents a very significant step forward when it comes to precise indoor positioning capabilities, promising to unlock game-changing products and services.

Advertisement

Thanks to global navigation satellite system (GNSS) technology, most of us now take the ability to pinpoint the location of an object or person outdoors for granted. The levels of precision positioning that can be achieved – particularly when you use a GNSS augmentation data service – are unlocking truly transformational use cases, including autonomous vehicles and other equipment.

Having had our eyes opened to the art of the possible outdoors, there’s now a growing desire to create similarly ground-breaking indoor applications.

One example is high-precision industrial- and consumer-grade asset tracking, or ‘find my item’ services. Another is indoor navigation, where someone (or something) is guided through a complex facility, such as a hospital or large transport interchange, with turn-by-turn navigation. A variation on this could see information displayed on a moving device when it reaches a particular point of interest. This could be to enable smart guidebooks in a museum or gallery, capable of providing information about a specific exhibit when someone approaches it.

To provide the sorts of user experiences their designers aspire to, applications such as these require a combination of highly accurate positioning indoors, low power consumption and affordable costs. This can’t be achieved with traditional short-range radio technologies.

The drawbacks of conventional Wi-Fi and Bluetooth positioning solutions

Wi-Fi Fingerprint, for example, only provides accuracy down to around 10m. Wi-Fi time-of-flight, also known as round-trip time (RTT) or 802.11mc, achieves 1-2m, but with high power consumption. While it has been available for some time, this technology has not yet been deployed in the field at scale.

Locating a device using Bluetooth, meanwhile, has traditionally used the ‘received signal strength indicator’ (RSSI). Based on RSSI, it’s possible to estimate the distance between a so-called ‘anchor point’ reference station and a device, though not the angle a device is at, relative to the anchor. By using three or more anchor points, you can then calculate the device’s position, but only with accuracy levels of between three and five meters.

Bluetooth direction finding: indoor positioning goes granular

Version 5.1 of the Bluetooth specification introduced ‘direction finding’. This represents a very significant step forward when it comes to precise indoor positioning capabilities – so much so, that it promises to unlock some of the game-changing products and services that today’s engineers are looking to create. Direction finding doesn’t rely on the strength of the signal emanating from a Bluetooth-enabled device to work out its location. Instead, it uses fixed anchor points with multi-antenna arrays to calculate the angle of the signal, using either the angle of arrival (AoA) or the angle of departure (AoD) technique. By calculating the angles of the signals coming from/to at least three anchor points, and identifying where they intersect, you can get a sub-meter-level location reading for nearby devices.

u-blox - indoor-positioning figure 1
Figure 1. By working out the angle of the signals reaching three fixed anchor points, and calculating where these intersect, you can identify where a device is located to within a meter.

Let’s look briefly at how these angles are calculated.

In an AoA setup, a mobile device sends out a Bluetooth direction finding signal. This reaches each antenna in an anchor point’s array with a slight phase shift relative to the other antennas in that anchor point. These phase differences can be used to calculate the signal’s angle of arrival, assuming the signal propagates a planar wave. AoA can be used for tracking, or for real-time location services (RTLS).

u-blox indoor-positioning figure 2
Figure 2. An angle of arrival (AoA) setup calculates the angle of an incoming Bluetooth direction finding signal based on slight phase differences observed at each antenna in a multi-antenna array.

With AoD, the signal is sent from each antenna on an anchor point to a nearby Bluetooth-enabled device. These signals reach that device with a slight phase shift. Coupled with information on the geometry of the antennae, these differences in phase can be used to calculate the signal’s angle of departure from the anchor. AoD is an effective way of implementing wayfinding and navigation solutions.

u-blox indoor-positioning figure 3
Figure 3. An angle of departure (AoD) setup sees the receiving device calculate the angle at which the signal left the antenna array, based on phase differences between signals sent from different antennas.

Putting it to the test

We’ve extensively tested the Bluetooth 5.1 direction finding technology in a number of use cases. A recent example was an AoA proof of concept, where we used a servo to create a rotating head that would turn to face the mobile device we were tracking. This would provide visible proof that the direction finding solution was working.

The mobile device in this particular setup was a u-blox application board with a NINA-B406 Bluetooth 5.1 low energy module on it. Its broadcast range was set to around 10m, though this can be extended if required.

In this POC, we had a single anchor point that included a u-blox antenna board with a NINA-B411 Bluetooth low energy module and our u-connectLocate software (more on this shortly). The antenna board contains five cross-polarized antenna elements, to determine the incoming signal angle in both the vertical and horizontal planes. The software on the board then calculates the angle of arrival. This information is then used to rotate the moving head to face the mobile device in real time.

The video below shows the setup in action, with the head continually tracking the device as it’s carried around the room.

We also did an extensive test in an industrial warehouse, which we describe in this indoor positioning white paper.

Enabling these algorithms in resource-constrained embedded systems

One of the most complex and time-consuming aspects of creating solutions using AoA or AoD is the implementation of the algorithms to calculate the signal angle from the phase differences, given the resource-constraints inherent in the embedded systems where these will typically be run.

This is why we developed the u-connectLocate direction-finding software for our NINA-B410 and B411 modules. This provides an API that enables engineers to obtain the AoA, handle data collection and pre-processing, and suppress multi-path components on each individual antenna. The software runs in Bluetooth module’s embedded MCU, meaning no external processing is needed. This reduces complexity and the bill of materials.

These are genuinely exciting times for anyone looking to create solutions that require highly precise indoor positioning capabilities. Bluetooth direction finding can unlock the sub-meter-level accuracy that opens the door to countless new use cases. Whether you’re looking to help people find their way through a hospital, pinpoint the exact location of a piece of machinery in a large warehouse complex, or do something else entirely, Bluetooth direction finding could be the game-changing technology you need.


Erik Carlberg - u-blox

Erik Carlberg is a senior product manager in the short range radio product center within u-blox AG. He is responsible for the roadmap of Bluetooth and Wi-Fi modules, targeting the industrial market.  Erik has 20 years of experience from the wireless industry. He holds a Master of Science degree in electrical engineering and a Bachelor in Business Administration, both from Lund University in Sweden.


Related Content:

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.