Measuring cortisol levels as a stress biomarker is essential in many medical conditions associated with a high risk of metabolic syndromes such as anxiety and cardiovascular diseases, among others. One technology that has a growing interest in recent years is fiber optic biosensors that enable ultrasensitive cortisol detection. Such interest is allied with progress being achieved in basic interrogation, accuracy improvements, and novel applications. The development of improved cortisol monitoring, with a simplified manufacturing process, high reproducibility, and low cost, are challenges that these sensing mechanisms still face, and for which solutions are still needed. In this paper, a comprehensive characterization of a D-shaped fiber optic immunosensor for cortisol detection based on surface plasmon resonance (SPR) enabled by gold coating is reported. Specifically, the sensor instrumentation and fabrication processes are discussed in detail, and a simulation with its complete mathematical formalism is also presented. Moreover, experimental cortisol detection tests were performed for a detection range of 0.01 to 100 ng/mL, attaining a logarithmic sensitivity of 0.65 ± 0.02 nm/log(ng/mL) with a limit of detection (LOD) of 1.46 ng/mL. Additionally, an investigation of signal processing is also discussed, with the main issues addressed in order to highlight the best way to extract the sensing information from the spectra measured with a D-shaped sensor.

This paper proposes an approach to process the response of a distributed temperature sensor using a nonlinear autoregressive with external input neural network. The developed model is composed of three steps: extraction of characteristics, regression, and reconstruction of the signal. Such an approach is robust because it does not require knowledge of the characteristics of the signal; it has a reduction of data to be processed, resulting in a low processing time, besides the simultaneous improvement of spatial resolution and temperature. We obtain total correction of the temperature resolution and spatial resolution of 5 cm of the sensor.

In this work, we analyze different types of recurrent neural networks (RNNs) working under several different parameters to best model the nonlinear optical dynamics of pulse propagation. Here we studied the propagation of picosecond and femtosecond pulses under distinct initial conditions going through 13 m of a highly nonlinear fiber and demonstrated the application of two RNNs returning error metrics such as normalized root mean squared error (NRMSE) as low as 9%. Those results were further extended for a dataset outside the initial pulse conditions used on the RNN training, and the best-proposed network was still able to achieve a NRMSE below 14%. We believe that this study can contribute to a better understanding of building RNNs employed for modeling nonlinear optical pulse propagation and of how the peak power and nonlinearity affect the prediction error.