(586c) AI-Driven Soft-Sensor Feedback Control for Microfluidic Production of Anisotropic Rods

The establishment of precise microscale environments through the formation of microfluidic
bubbles and droplets is crucial for advancements in fields such as pharmaceuticals, chemical
synthesis, and DNA sequencing. Achieving precise control over the size, shape, and
functionality of these entities is essential for their effectiveness over emulsions produced using
bulk fabrication methods. Despite efforts to maintain constant external variables such as flow
rates and pressures, unpredictable factors can disrupt microfluidic processes, jeopardizing the
uniformity of resulting emulsions. In this study, we introduce a two-step soft-sensor approach
incorporating a convolutional neural network (CNN) and an image recognition algorithm for
feature extraction of the size and aspect ratio of anisotropic hydrogel rods used in tissue repair
studies. This method facilitates the detection of flow regimes and enables assessment of
emulsion size and aspect ratio, and when integrated with a proportional-integral-derivative (PID)
controller, the soft sensor demonstrates effective setpoint tracking of the size and aspect ratio of
the hydrogel rods. Through the utilization of the soft-sensor and AI-driven feedback control, our
research presents a widely applicable methodology for precise and automated microfluidic
control across diverse applications.