Project BRAINMIX

Brain-inspired unsupervised multimodal learning method for fast motion measurement by self-mixing interferometry

About the project

Research at the LEAT has demonstrated how the fusion of sensory modalities can enhance the quality of signal classification or reconstruction in noisy environments by using brain-inspired learning methods. These methods are particularly well-suited to applications in optical feedback interferometry, a nonlinear optical measurement technique with a broad range of detection applications. However, its practical use remains challenging due to the complex data analysis it requires and its sensitivity to tiny variations in the target's reflectivity. Recent research conducted at INPHYNI has shown that, while a convolutional neural network can largely address the first issue, it cannot resolve the second. Leveraging the expertise of both teams, we will develop bio-inspired multimodal learning methods in the context of a sub-micrometer displacement measurement system based on optical feedback interferometry modalities.