Multi-source gateway for collecting air quality data from a smart-building

Multi-source gateway for collecting air quality data from a smart-building

The goal of the MARINA R&D project is to propose optimal ventilation management of tertiary buildings. In this project, CETIC brings its expertise in the advanced and transparent management of the sensors’ heterogeneity and their data. Test sites are deployed, including among other innovative Volatile Organic Compound (CoV) sensors. This will ultimately make it possible to offer a customizable innovative facility management service.

Date: 5 April 2023


Scalability of embedded systems and IoT networks 

Domaine: Energy and environment 

Asset: Dmway 

About project: Marina 

Author : Amel Achour

In the MARINA project, CETIC provides expertise in the advanced and transparent management of the heterogeneity of the sensors involved and the data they generate (see Figure 1). CETIC is developing a middleware, DMWAY (Dynamic Mapping Gateway), allowing efficient management of IoT data flows from/to heterogeneous data sources, interfaces and protocols, thanks to its modular, extensible and configurable architecture through the data box specialized module tools available to DMWay (see Figure 2).

Figure 1. DMWAY universal Gateway
Figure 2. DMWay Toolbox

Figure 3 represents the DMWay toolbox’s use for the implementation of the CETIC building monitoring demonstrator with, as input, the data sources currently supported and, as output, the connectors to the remote systems currently implemented.

Figure 3. Architecture of CETIC building monitoring

In the MARINA project, the support and integration of air quality data sources is done incrementally through the implementation of Proofs of Concept (PoC). This enriches the interfaces and protocols supported by DMWay. Two PoCs are deployed, with the partners BizzDev and VOCSens.

Deployment of DMWAY at the BizzDev partner site

The objective of this PoC is to evaluate DMWAY in the BizzDev partner’s environment. The PoC is represented in the figure below, it allows the collection of data from connected sockets and multi-sensors deployed in the premises of BizzDev, and from a remote sensor of ammonia - NH3 (developed by the VOCSens partner for this example) deployed in CETIC premises. The latter sends the measurements remotely to the BizzDev infrastructure using the RabbitMQ connector. Thus, this PoC illustrates DMWay’s ability to manage various data sources, and to connect effectively to the processing and visualization service from different locations (Gosselies and Tournai in this example).

Figure 4. Illustration of the BizzDev PoC

The figure below represents the interfaces and connectors integrated into DMWAY for the implementation of this PoC.

Figure 5. DMWAY components of the BizzDev PoC

The figure below shows the visualization dashboards at the two sites.

Figure 7. Visualization Dashboards

VOCSens Sensor integration with own semantics and data format

The objective of this PoC is to implement the interface allowing the VOCSens sensor’s support. This sensor communicates through a serial link. It is connected to a RaspberryPi on which a DMWay is deployed. A Mapper is implemented to translate the data received from the sensor in JSON format to a datamodel associated with it internally in DMWay, this work is done only once. Thus, the integration of new sensors with the same characteristics is greatly facilitated thereafter:

  • If the new sensor has the same functionality, no further action is required for it to be supported.
  • If the new sensor has new features, these will be integrated into the existing plugin and will be natively supported for future sensors.

This PoC also allowed us to develop and validate connectors for the simultaneous transmission of data to two different applications, so the data collected is used (see Figure below):

  • locally and viewed at CETIC through the local ThingsBoard IoT platform instance.
  • remotely, at the BizzDev infrastructure through the RabbitMQ broker
DMWAY components of the VOCSens PoC


Through these two PoCs, we showed the flexibility of DMWay to manage different sensors with heterogeneous data formats. This is easily and quickly customizable thanks to the mechanisms offered by DMWay to support the specific semantics of each data source, whether it comes from an established standard or from a proprietary definition. We also demonstrated the ability of DMWay to transmit the data collected transparently to different remote systems.
The positive results at these first stages of the project make it possible to continue towards the development of an optimized service for indoor air quality management in buildings in the tertiary sector.