Gas Detection

Gaussian Process Models and Reaction Signatures

Interpreting Gas Detections

Our client was a start-up who have innovative technology for gas detection based on precise doping of materials to react with passing gases. They had produced a range of reactive compounds and a significant volume of data from controlled reactions with individual and mixtures of gases, when they bought React AI on board to assist with interpreting this data.

React AI spent some time working closely with our client’s team, before leaving them with the tools, techniques and understanding they needed to move their process forward.

Services Provided

React AI undertook a collaborative role with our client to assist them in understanding their data, including:

Review of the processing pipeline for data collection and analytics.
Refactoring and extension of the pipeline.
Analysis of gas reaction readings vs. detector composition, including using Gaussian process models.
Extraction of data from noisy reaction curves, with neural network and other techniques.
Analysis across the data set of gas reactivity ‘signatures’, to establish how to maximise accuracy for desired detections while minimising the different detection compounds required; combination of this with the Gaussian models to produce tools to predict the levels of doping required to produce the optimum compounds for specific tasks.
Provision of processing-pipeline and analytics code, with user guide and documentation.
General advice and consultancy; reports on work undertaken, documentation of all results, conclusions and recommendations for future stages.

Contact Us

Gas Detection

Interpreting Gas Detections

Services Provided