We developed a lightweight 1D Convolutional Neural Network (CNN) model optimized for edge devices, reducing inference latency while maintaining high classification accuracy. Using an open-source dataset of 30 LoRa devices, we evaluated preprocessing methods (Short Time Fourier Transform (STFT) and Fast Fourier Transform (FFT)). Our approach demonstrated: A significant reduction in inference latency, making deployment on real-time edge devices feasible. Comparable classification accuracy when benchmarked against a 2D CNN model. This work not only bridges a crucial gap in the literature but also propels the adoption of RFF for edge devices—an essential step for secure IoT networks in 6G. By aligning this effort with COST Action, we aim to foster collaboration across Europe to achieve secure, resilient, and trustworthy 6G systems.