Introduction

Bluetooth interface implementation in RouterOS allows the device to capture Bluetooth advertising packets that are broadcasted over 37, 38, and 39 advertising channels. More information can be found in the guide here.

Bluetooth tags like, for example, TG-BT5-IN and TG-BT5-OUT, do exactly that. They broadcast advertising payloads over the mentioned channels. To understand what kind of information is stored in the payload, make sure to check the link. The tags can be configured (using the MikroTik Beacon Manager app) to broadcast the payloads automatically, with an interval or/and when a movement, tilt, or free-fall trigger is detected. In simple terms, the tag will "tell" (broadcast to) all the surrounding Bluetooth scanners (like the KNOT) information about itself periodically.

When the payload is broadcasted by the tag, and the tag is within the KNOT's Bluetooth operating range, the KNOT will capture and display the payload under its "scanner" Bluetooth interface section. It would look like this:

/iot bluetooth scanners advertisements print
Columns: DEVICE, PDU-TYPE, TIME, ADDRESS-TYPE, ADDRESS, RSSI, LENGTH, DATA
#  DEVICE  PDU-TYPE        TIME                  ADDRESS-TYPE  ADDRESS            RSSI    LENGTH  DATA                                        
0  bt1     adv-noconn-ind  mar/07/2023 12:11:57  public        DC:2C:6E:0F:C0:3D  -51dBm      22  15ff4f09010079100000ffff0000cf188a6b2b000064
1  bt1     adv-noconn-ind  mar/07/2023 12:11:58  public        2C:C8:1B:4B:BB:0A  -49dBm      22  15ff4f090100168dfefffffffeffa51ae1362200005e 

The example above shows us, that the KNOT sees two Bluetooth tags within its operating range with MAC addresses "DC:2C:6E:0F:C0:3D" and "2C:C8:1B:4B:BB:0A", their respective payloads ("DATA" field) and the signal strength ("RSSI" field).

With the help of RouterOS scripting and scheduling, we can make the KNOT automatically-periodically scan the payload list and, in case, a specific payload or a specific tag's MAC address is found on the list, we can make the KNOT structure an MQTT message (out of the printed information shown in the example above) and send it to the configured server via MQTT, e-mail or HTTP post. Script examples will be shown later on in the guide.

As the title suggests, the goal is to implement a Bluetooth tag-tracking solution and the idea is quite simple. When you have 2 KNOTs (KNOT-A and KNOT-B), running the same script on a scheduler, and the tag moves between their Bluetooth operating ranges, the data on the server will indicate whether it was KNOT-A or KNOT-B that have sent the tag's payload. That will help you figure out the proximity of the tag. Whether the tag is broadcasting payloads in the KNOT-A zone, or in the KNOT-B zone.

You will need a server where the data is going to be stored and visualized. In this guide, we will showcase a server called ThingsBoard and how to communicate with it using the MQTT protocol.

ThingsBoard has a cloud solution and different local installation options (on different OS). Since we've added a container feature, it became possible to also run the platform within RouterOS. In order to do that, you will need a RouterOS device that has at least 2 GB RAM, has an option to increase storage (for example, an additional USB port), and is either ARM64 or AMD64 architecture. CHR machine or CCR2004-16G-2S+ could be a good fit.

Setup requirements:

• a running ThingsBoard server;
• 2+ KNOTs with access to the server's network via ethernet, Wi-Fi, or cellular connection (the amount of the units required depends on the size of the area that needs to be covered);
• 1+ Bluetooth TG-BT5-IN and/or TG-BT5-OUT tags (depending on how many assets you need to track - 1 tag per asset).

Scenario explanation

Imagine LTE base station roaming or Wi-Fi roaming topologies.

In the LTE case, you have multiple base stations that cover specific areas in your city. In populated areas, they are scattered around so that they cover the whole city. You have a smartphone with an LTE connection up. When your stay at home, you are most likely connected to a single base station (not always but most of the time), which provides the best signal. ISP knows that your phone is connected to a specific base station, for example, base station #5. When you go to work, your phone will roam between the base stations on the way, and once you arrive at your office, your phone will, most likely, be connected to another base station, for example, #10, because the base station #5 will be out of reach. Now, your ISP knows that your device is connected to station #10.

In the Wi-Fi roaming case, you have multiple access points broadcasting the same SSID. When you stay at your workstation, you will be, most likely, connected to AP #1. Then, if you go downstairs to visit your coworker on another floor, you will be roamed to AP#2, as AP#1 will be out of reach.

The same principle can be applied to Bluetooth tag-tracking. Instead of the "LTE base stations" or "Wi-Fi access points", you will have the KNOTs (Bluetooth scanners) and instead of the smartphone that roams between the base stations or access points, you will have the Bluetooth tag that roams between the KNOTs.

The actual Bluetooth operating distance can vary from site to site because a lot of different factors can impact it, like 2.4 GHz interference or the surrounding materials used. For example, in line of sight, with no interference, the distance at which the KNOT is able to capture the tag's broadcasted payload, can be up to 180 meters (KNOT — ~180 meters — TG-BT5-OUT). But you also have to keep in mind that with further distance, more packets will be lost on the way. In the office environment, the range can go down to 30-100 meters.

Let's take a look at a basic-rough example first. We have two KNOTs (KNOT-A and KNOT-B). We've tested Bluetooth ranges in our environment and could verify that both KNOTs are able to capture the tag at a 70-meter distance. If we install KNOT-A and KNOT-B 140 meters apart, their Bluetooth ranges will not overlap or overlap just slightly. When the tag moves into the KNOT-A range, the payload from the monitored tag will appear under the Bluetooth scanner list, an MQTT message with a report will be sent to the server and the server will display that the tag is in the KNOT-A zone. When the tag moves into the KNOT-B range, the same happens, and the server will display that the tag is inside the KNOT-B zone.

Logically, if the Bluetooth operating ranges overlap and the tag is within the overlapped area (at the same time within KNOT-A and KNOT-B Bluetooth ranges), both KNOTs will send the data and the server is going to show that the tag is reported by both devices at the same time.

Surely, there will be zones where two or more KNOT's Bluetooth ranges overlap and you can use it to your advantage → basically, you will have information that the tag, right now, is on the edges of the Bluetooth ranges in-between specific KNOT zones. In other words, when the asset moves into the overlapped area, you will have information on the server that the asset is somewhere between the two KNOT operating ranges, which is useful information to have as it gives more precision.

Additionally, the script we've prepared allows you to set up a filter (that will be shown later on), which will make the KNOT ignore the payloads that the scanner captures unless the signal strength (RSSI) is stronger than the specified value. In the Bluetooth scanner example print shown above under the "Introduction" section, we can see that the KNOT sees one of the tags with an RSSI signal strength of -51 dBm (tag with a MAC address "DC:2C:6E:0F:C0:3D") and the other one with an RSSI signal strength of -49 dBm (tag with a MAC address "2C:C8:1B:4B:BB:0A"). So, if we apply a filter to the script to ignore all payloads that are received with a signal strength (RSSI) weaker than -50 dBm, our KNOT would report that only tag "2C:C8:1B:4B:BB:0A" is within the Bluetooth range, because its RSSI is -49 dBm, and the second tag (with RSSI -51 dBm) will be ignored.

Bluetooth advertisements

    # whose signal strength is stronger than -40dBm

To understand the scenario and the idea behind it better, take a look at the topology shown below:

Let's say you have a warehouse and you wish to track pallets. Just install the KNOTs in different locations

Configuration

/iot/mqtt/brokers/add name=tb address=x.x.x.x port=1883 username=access_token