Passive indoor position meter

Project Objective: Development of a Prototype System for Collecting Passive Indoor Visitor Statistics and Position Data

The objective of our project was to develop a prototype system that enables the position of visitors to be determined in large buildings under indoor conditions, based on their WiFi and/or Bluetooth-enabled devices. Receiver stations placed at specific points within the monitored area observe standard WiFi and Bluetooth communication and list identifiable mobile devices and portable computers, while the data collection software system generates position data from simultaneous point-based observations. In coordination with the visitor counting systems manufactured and distributed by our company, the product offers our clients innovative statistical functions for visitor tracking and visitor behaviour analysis, including the description of movement within buildings.

The hardware component of our system consists of commercially available devices, supporting simple and cost-effective procurement, as well as the replacement and modernisation of deployed systems. Counting and tracking data are displayed through a web-based interface accessible from a browser, both in tabular statistical form and visually. The relative traffic of individual positions is represented on a heat map, which is one of the most illustrative ways to display this type of data.

Solution Architecture

The system has a modular structure and consists of two main logical components:

• Primary data collection is performed by receiver units (“receivers”, “antennas”) assembled from commercially available hardware components. These must be installed in a way that provides multiple coverage of all parts of the monitored area.
• In parallel, data processing is carried out by specialised central software solutions developed within the project. By comparing and aggregating the data from the individual receivers, these software components are also capable of determining the spatial characteristics of visitor data.

For technical and data protection reasons, there is also an intermediate processing layer implemented on the receiver-unit side. This layer anonymises the incoming data for GDPR-related reasons and enriches it with timestamps and metadata required for spatial reconstruction.

The modularity of the system is based on the fact that the number of receivers can be scaled according to the characteristics of the deployment site. A single server can manage receiver units of any type and in any number, up to the limits of its computational capacity.

The receiver units are available in two versions. Within the scope of our project, we developed both a shorter-range island receiver and a longer-range perimeter receiver. Based on our experience, island receivers may provide sufficient data on their own for calculating position data; however, due to their proximity-based operation, an adequate number of such units must be installed at every location. Perimeter receivers are higher-performance units equipped with directional antennas. They are suitable for further refining positions, but on their own they are not sufficient for average-sized or larger sites.

A relational database, application server components, and custom-developed software components run on the central server unit, above a Linux-based operating system. Together, these components provide the system’s position measurement functionality. This central unit monitors and coordinates the operation of the receiver units and receives raw reception data from them. Based on this data, our proprietary triangulation and position calculation algorithms provide the calculated map positions.

The system includes a user interface running above the application server, which is capable of displaying visitor and position data, including heat maps. The user interface is an online, browser-accessible interface that complies with modern standards while also ensuring compatibility with our company’s conventional camera-based visitor counting systems.

Technological Background

The theoretical basis for the operation of the system is the fact that today almost every visitor owns and carries one or more mobile devices with WiFi or Bluetooth connectivity switched on, including phones, tablets, and even smartwatches or earphones. During operation, these devices transmit packets over open radio channels, in a manner that can be received by anyone. Although the content of these packets is encrypted and irrelevant to us, the packets themselves contain device-specific identification data. By collecting radio packets transmitted within a given area, and by systematically collecting and organising device-specific identifier tags, we can obtain information about nearby devices and, under suitable conditions, determine their position.

In the case of WiFi, the operation can be illustrated as follows: when WiFi is switched on, devices transmit so-called “probe requests” even without an active connection. These are used by the devices in their background operation to list available WiFi networks. Our system collects these packets passively. The MAC address of the device, which serves as the primary identifier, is typically present in the individual packets. In some cases, the names of the WiFi networks searched for by the device are also present. While these names are not informative individually, their combination can be considered a device-specific identifier. On this basis, packets originating from the same device can be recognised.

Positioning also requires the recorded signal strength, or RSSI, for each packet, which became the basis of the triangulation procedures used by our island receivers. As a counterpart to this, in the case of perimeter receivers, the receiving antenna is also used. In our project, we observe signals at the 2.4 GHz frequency, because at 5 GHz there are 165 possible channels instead of 14, which could only be observed using significantly more complex hardware or with less favourable resolution in a time-division-based approach.

Results

Within the scope of our project, we developed a prototype system providing the targeted functionality. We assessed the range of applicable radio devices required for system deployment, established the software framework of the system, and created the specialised algorithms that provide the custom functionality. As part of the project, we deployed a pilot system at a large-area location with a business function, and summarised the deployment and operational experience in the form of a case study.

In order to disseminate our results widely, we participated in the relevant conference of the Hungarian Council of Shopping Centres and presented the capabilities of our system and the results achieved to our professional partners.

To support broad dissemination, we are publishing the Windows desktop version of the framework software implementing the system as free-of-charge software, which can be used for an unlimited period of time.