This section summarizes the main deliverables, scientific results and dissemination outputs of the NEURALSENS project.
The project combines the development of hydrogen and temperature sensing elements, Sensory Neural Network (SeNN) design, in-sensor computing concepts and the implementation of smart sensor architectures based on crossbar arrays.
Project deliverables
D0.1 — Project website
Public website providing information about the NEURALSENS project, its objectives, partners, progress and results.
D1.1 — International conference presentation
Presentation of results related to the development and characterization of H₂ sensing elements.
D1.2 — Paper on H₂ sensing element
Scientific publication focused on the development and characterization of the hydrogen sensing element.
D0.2 — Project interim report
Interim report summarizing project progress, achieved results and the status of individual work packages.
D2.1 — International conference presentation
Presentation of results related to Sensory Neural Networks, sensor response processing and in-sensor computing.
D2.2 — Paper on smart temperature sensor
Scientific publication describing the smart temperature sensor and the validation of the SeNN design approach.
D3.1 — International conference presentation
Presentation of results related to the development of the SeNN laboratory prototype.
D2.3 — Open-source SeNN simulation software
Open-source software framework for simulation and design of Sensory Neural Networks based on crossbar-compatible resistive sensing elements.
D0.3 — Final project report
Final report summarizing the scientific and technical outcomes of the NEURALSENS project.
D3.2 — Paper on H₂-sensing CBA SeNN laboratory prototype
Scientific publication describing the laboratory prototype of the hydrogen-sensing crossbar-array Sensory Neural Network.
Selected publicly available project outputs
NEURALSENS project website
The NEURALSENS website provides public information about the project concept, objectives, partners, review meetings, deliverables and selected scientific results.
Project website: neuralsens.org
NEURALSENS SeNN simulation framework
A Python-based simulation framework for Sensory Neural Networks has been developed within the project and is publicly available through GitHub.
The framework explores the use of sensor characteristics as physically defined network parameters and provides tools for modelling and comparing conventional neural-network processing with sensor-dependent SeNN architectures.
GitHub repository:
https://github.com/jaro221/NEURALSENS_PNN
IEEE NAP 2026
Geometry-Driven Temporal Signal Encoding in TiO₂ Gas Sensors for Neural Network Integration
P. Nemec, J. Klarák, M. Predanocy, M. Horský, J. Škriniarová and B. Hudec
Presented at the IEEE International Conference on Nanomaterials: Applications & Properties (IEEE NAP 2026), Budva, Montenegro. Download conference poster (PDF).
The work demonstrates that controlled sensor geometry modifies both the magnitude and temporal evolution of the hydrogen response. At 1000 ppm H₂, the investigated geometries exhibited a response-factor range spanning more than two orders of magnitude. Temporal response features were further used to construct an NN-ready representation of the sensor signal.
IEEE NAP 2025
Application of a TiO₂ Gas Sensor as a Component in Neural Networks
The contribution investigated the use of geometrically diverse TiO₂ gas-sensing elements as inputs to neural-network-based processing and established an initial link between sensor design and neural-network integration.
Ongoing and final outputs
Additional publications, conference contributions and laboratory-prototype results will be added as the corresponding project outputs are finalized and made publicly available.