hAIdrogen safety sensors

Important results from the project

The project has achieved its goals by proving that AI can be used to (i) accelerate H2 sensors up to 40 times; (ii) enable H2 sensors to operate in chemically demanding environments, such as high humidity in air; (iii) process data generated by a multiplexed sensor prototype with up to eight different active materials, (iv) handle drift in raw sensor data. In addition, we have shown that AI-multiplexed sensors can measure other species than H2 simultaneously with high precision.

Expected long term effects

The project´s results have long-term effects due to the groundbreaking nature of the results, which we expect to open up new research directions where others in the sensor field will be inspired and follow suit. This, in turn, will accelerate the use of AI in sensor science and technology in general, as well as the development of multiplexed sensors for complex measurement challenges outside the project´s focus of H2 sensors.

Approach and implementation

The project was implemented as a collaboration between two academic partners and an SME that develops H2 sensors. The academic partners were experts in H2 sensors and AI. The SME´s role was to provide advice on the technical relevance of the research and the development of a prototype. The setup worked well and the project developed as intended. The schedule could not be followed completely as planned due to delayed recruitment of a postdoc and parental leave of another key person in the team.