In a realistic odorant detection application environment, the collected sensory data is a mix of unknown chemicals with unknown concentrations and noise. The identification of the odorants among these mixtures is a challenge in data recognition. In addition, deriving their individual concentrations in the mix is also a challenge.
A deterministic analytical model was developed to accurately identify odorants and calculate their concentrations in a mixture with noisy data. This model is specially suited for hardware implementation with miniaturization. Hierarchical neural network architecture effectively deals with the induced odorants that can be formed from the combination of basic source odorants and their concentrations.
To search for an odorant in the mixture, where it exists in the operating environment, one of the most robust techniques is to recover the original odorant sources. When done, the detection can be an easy step by finding the minimum phase between the predicted original odorants and the target odorants. The neural-network approach can be employed to capture the target odorants in various conditions through learning, i.e., concentration levels through the parameterized weight set, then the strongest correlation between parameterized weights and the predicted original can be used to identify the intended odorants.