報告題目：Next generation optical fibre sensors powered by machine learning
報 告 人：Dr Stephen Warren-Smith
邀 請 人：智能檢測與控制系 李雪剛、張亞男
The propagation of coherent light in multimode optical fibers results in complex light patterns known as speckle. This effect has historically made understanding and controlling the light transport in these fibers extremely difficult, something which has only recently been solved for applications such as imaging. This complexity can also be a powerful tool for sensing environmental parameters, as small perturbations lead to significant changes in the output of the fiber. In this talk I will explain how deep learning can be utilized to extract the rich sensing information embedded within the multimode optical fiber output and highlight applications that can benefit from this approach. This includes multipoint high-temperature sensing using sapphire crystal fiber, pressure sensing, and motion sensing for healthcare applications.
Stephen is an ARC Future Fellow (澳大利亞杰青) and leader of the Intelligent Photonics Lab within the Future Industries Institute at the University of South Australia. His research focuses on creating fundamental developments in optical physics, data science, materials science, and fabrication and sensing technology, and apply them to critically needed challenges areas including biomedical applications, health care, and heavy industry. His latest efforts focus on intelligently exploiting the complexity of multimode photonics to create the next generation of sensors. Stephen is passionate about promoting both science to the wider community, including his current position as Chair of the South Australian Branch of the Australian Institute of Physics.