Enhancing GNSS Signal Integrity in Medical Logistics: A Deep Learning Solution

Introduction: Ensuring the integrity of Global Navigation Satellite System (GNSS) signals is critical for the timely and accurate delivery of pharmaceuticals within smart medical supply chain logistics (SMSCL). Method: In this study, we propose a novel deep learning (DL) framework that integrates a Bidirectional Long Short-Term Memory (BiLSTM)-Attention model with Principal Component Analysis (PCA) and Bayesian Optimization (BO) for feature selection. This approach enhances GNSS signal reliability by accurately detecting anomalies, especially in environments prone to interference. The PCA-BO feature selection process optimizes relevant features like signal strength and Doppler shifts, improving model performance while reducing overfitting. Results and Discussion: Our results demonstrate that the proposed model significantly outperforms conventional methods, enhancing the precision of pharmaceutical deliveries in critical healthcare settings. Conclusion: This work represents a key advancement in using DL to ensure GNSS signal integrity for SMSCL, contributing to more efficient and secure logistics operations.