E-Nose and e-Tongue combination for improved recognition of fruit juice samples

Area of Science:

• Analytical Chemistry
• Food Science
• Sensor Technology

Background:

• Food security analysis faces challenges with current sensory evaluation tools.
• Reliable methods are needed for comprehensive analysis of smell, taste, and color.
• Multisensor systems offer potential for integrated sensory data.

Purpose of the Study:

• To investigate a multisensor data fusion approach combining an electronic-nose (e-Nose) and an electronic-tongue (e-Tongue).
• To classify Tunisian fruit juices using a low-level data fusion technique.
• To evaluate the performance of combined e-Nose and e-Tongue data against individual systems.

Main Methods:

• Utilized five tin oxide-based Taguchi Gas Sensors for the e-Nose.
• Employed six potentiometric sensors for the e-Tongue.
• Applied Principal Component Analysis (PCA) and Fuzzy ARTMAP neural network for data analysis and classification.
• Characterized eleven fruit juice varieties from four commercial brands.

Main Results:

• Individual e-Nose and e-Tongue analyses using PCA showed limited discrimination between fruit juices.
• The low-level data fusion technique combined with Fuzzy ARTMAP achieved 100% success rate in fruit juice recognition.
• The data fusion approach provided complementary and comprehensive information, enhancing classification accuracy.

Conclusions:

• Multisensor data fusion effectively merges data from multiple sources for superior analytical outcomes.
• Combining e-Nose and e-Tongue signals offers a powerful strategy for detailed fruit juice characterization.
• This approach significantly outperforms the performance of individual sensor instruments in food analysis.