This paper addresses the challenges of data privacy and communication efficiency in mobile telemedicine systems, particularly in the context of 5G-and-beyond networks and federated learning (FL). The authors propose an adaptive scheduling mechanism to reduce communication loads and an interpretable convolutional fuzzy rough neural network (CFRNN) to enhance model interpretability. They also develop a multiobjective memetic evolutionary algorithm (MOEA) to optimize neural network architectures. The framework integrates smart scheduling, interpretable fuzzy rough logic, and neuroevolution, aiming to improve communication efficiency, increase diagnostic interpretability, and protect patient privacy. The contributions include an adaptive communication scheduling scheme, a CFRNN model, and a flexible neural architecture search space optimized by a MOEA. The paper is structured into sections covering preliminaries, the proposed framework, experimental details, and conclusions.This paper addresses the challenges of data privacy and communication efficiency in mobile telemedicine systems, particularly in the context of 5G-and-beyond networks and federated learning (FL). The authors propose an adaptive scheduling mechanism to reduce communication loads and an interpretable convolutional fuzzy rough neural network (CFRNN) to enhance model interpretability. They also develop a multiobjective memetic evolutionary algorithm (MOEA) to optimize neural network architectures. The framework integrates smart scheduling, interpretable fuzzy rough logic, and neuroevolution, aiming to improve communication efficiency, increase diagnostic interpretability, and protect patient privacy. The contributions include an adaptive communication scheduling scheme, a CFRNN model, and a flexible neural architecture search space optimized by a MOEA. The paper is structured into sections covering preliminaries, the proposed framework, experimental details, and conclusions.