AI and Edge Computing at the service of IoT : Realization of an Augmented Sensor

Within the framework of a research & innovation working partnership on AI (Artificial Intelligence) and Edge Computing, the Adeunis and CARL/Berger-Levrault R&D teams have been working together for several months. CARL/Berger-Levrault is a European leader in equipment management (CMMS/EAM) and technical asset management, while Adeunis is a specialist in radio IoT solutions. These two entities are respectively located at the two ends of the IoT data value chain. This partnership is based on a common conviction: information processing must be split between Cloud processing and processing embedded in the physical sensor. It is this shared vision of the market, between two entities with complementary levels of knowledge, that led them to collaborate on this project.

Objectives and Challenges :

Initially, both entities chose to rely on the Adeunis Delta P sensor. This sensor monitors the proper functioning of the ventilation systems. The objectives and challenges are :

  • Embedded intelligence in Adeunis connected objects: anticipating system maintenance and better knowing the failures detected without processing in the Cloud, this reduces network bandwidth consumption and allows for local multi-sensor intelligence (data correlation). The sensor will be able to directly issue recommendations and alerts to CARL/Berger-Levrault’s equipment management solutions. The sensor can operate on battery power (1 to 10 years autonomy), and therefore requires no wiring for power supply or communication. Thus, we must meet the challenge of very low energy consumption for this connected object by optimizing our algorithms, models and via dynamic adaptation of data collection frequencies.
  • Facilitating the implementation of CARL Software’s AI algorithms on these products. The aim is to prevent and anticipate the technical maintenance of equipment by generating predictive models; these predictive models will be transmitted to the sensors, which will be able to instantiate them locally. The challenge will be to optimize and compress these predictive models to enable embedded execution on the connected object: low computing resources and minimizing the energy consumption of the connected object.

For the development of a global IoT solution, in addition to the use of AI, both entities also use Edge computing. Edge Computing is an optimization method used in cloud computing that consists of processing data at the edge of the network, close to the data source. This makes it possible to minimize the bandwidth requirements between sensors and data processing centers by undertaking analyses as close to the data sources as possible. This method is becoming essential for IoT. Edge Computing, by considerably reducing the volumes of data that transit, makes IoT solutions more efficient and more economical. It reduces latency and associated costs, improves security and speeds up decision making.

For the development of a global IoT solution, in addition to the use of AI, both entities also use Edge computing. Edge Computing is an optimization method used in cloud computing that consists of processing data at the edge of the network, close to the data source. This makes it possible to minimize the bandwidth requirements between sensors and data processing centers by undertaking analyses as close to the data sources as possible. This method is becoming essential for IoT. Edge Computing, by considerably reducing the volumes of data that transit, makes IoT solutions more efficient and more economical. It reduces latency and associated costs, improves security and speeds up decision making.

The project :

The project is to develop a global, universal, intelligent and low-cost IoT solution (economically adapted to the scale of a building) from the sensor to the visualization platform, for the predictive maintenance of building equipment. Ultimately, the two entities wish to simplify the life of their customers by removing technical and economic barriers. This integrated offer will provide an autonomous predictive maintenance service, adapted to each machine, simple to manage and deploy.
The ultimate goal is to offer more and more value-added services to their joint customers while remaining at the forefront of innovation. The advantages of this solution integrating AI and EC (Edge Computing) :

  • More accurate/relevant information provided
  • Anticipation of needs and increased reactivity for the maintainer
  • Better targeted maintenance actions
  • Reduction of intervention costs
  • Energy performance, maintenance costs and durability of equipment

More ...

Scroll to Top