Federated K-Means Clustering for Adaptive OFDM-IM

In this letter, a new federated learning strategy for k-means clustering assisted adaptive orthogonal frequency division multiplexing with index modulation (OFDM-IM) system, which develops a global model using local learning outcomes aggregated from distributed devices, is proposed. The proposed strategy aims to efficiently leverage the computing power of all the devices in a distributed system. Simulation results show that the proposed strategy reduces the training steps required at each device and simultaneously improve the throughput, with a federation cost for model aggregation and broadcast.