Artificial Neural Network-Assisted Confinement Loss Prediction of D-Shaped PCF-SPR Biosensor

Machine learning (ML) has become a pivotal force in enhancing the capabilities of sensing technologies across a wide range of applications, including biosensing, agriculture, industrial automation, and more. This paper presents a photonic crystal fiber (PCF)-surface plasmon resonance (SPR) biosensor with a maximum wavelength sensitivity (WS) of 41,000 nm/RIU and attains the highest resolution of 2.44 × 10<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">−6</sup> within an analyte refractive index (RI) of 1.39. The sensor’s highest sensitivity and resolution stem from its thoughtful design and the strategic choice of gold as a plasmonic material for the PCF, enhancing light-analyte interactions. Moreover, an artificial neural network (ANN) is used to accurately predict confinement loss with high precision and refine the biosensor’s design for optimal performance. Furthermore, a pivotal part of this research is to study various ML activation functions such as rectified linear activation (ReLU), hyperbolic tangent (Tanh), and Sigmoid to understand their role in accurately predicting confinement loss. Additionally, the work detects subtle changes in the RI of analytes holding various types of human cancerous cells. This advancement in SPR sensor technology heralds exciting developments in sensitive detection platforms, offering a powerful instrument for cutting-edge research.