Post

Aucxis goes all out for Machine Learning with the 'Smart RFID Gate' innovation project

Research project to distinguish tag reads of goods passing through the gate (item level) at a loading dock from waiting or passing goods.
News
Machine Learning - RFID Smart Gate

RFID gates with a reading range that is too high will be faced with incorrect or unwanted detections. In order to prevent this problem, we want to implement machine learning algorithms in a ‘Smart RFID gate’ that make it possible to distinguish tag reads of goods passing through the gate (item level) at a loading dock from waiting or passing goods.

Aucxis initiates this research project in collaboration with the company Boltzmann, specialising in research and development of machine learning applications. Concretely, we want to make use of the captured data – in which we detect patterns by means of algorithms – to improve our RFID gates’ performance. In this way, the gates can automatically learn, as it were, from their own experiences.

Benjamin Vandermarliere, Boltzmann: “The resulting algorithms must perform with high accuracy in a variety of conditions. This requires the necessary research, but I am confident that we can handle this problem in cooperation with the specialists at Aucxis”.

Data-driven machine learning algorithms will be used to to distinguish stray reads – unwanted reads of stray tags of waiting or passing goods – from desired reads with our Smart Gate. To be interesting for our customers, the allowed margin of error is very small.

Advantages: considerably less free space needed in front of and next to the gate, no need for a complex gate infrastructure with expensive shielding and absorption material such as cages, fences, RF absorbers,… to prevent stray reads, which implies obviously a significant cost reduction.

With this development project, Aucxis wants to improve the current market supply and develop new applications based on smart RFID technology. A major advantage of Machine Learning techniques is that one trained algorithm can be integrated into multiple systems.

This project is supported by the Flemish Agency for Innovation and Entrepreneurship (VLAIO).

This site is registered on wpml.org as a development site.