Add to Chrome
Add to Firefox
Add to EdgeAccumulated Polar Feature based Deep Learning with Channel Compensation Mechanism for Efficient Automatic Modulation Classification under Time varying Channels
Jan 06, 2020
In next-generation communications, massive machine-type communications (MTC) induce severe burden on base stations. To address such issue, automatic modulation classification (AMC) can reduce signaling overhead through blindly recognizing the modulation types without handshaking, thus playing an important role in intelligent modems. The emerging deep learning (DL) technique stores intelligence in the network, resulting in superior performance over traditional approaches. However, DL-based approaches suffer from computational complexity and offline training overhead, which severely hinder practical applications. Furthermore, the required massive resources for online retraining should be further considered under time-varying fading channel, which has not been studied in the prior arts. In this work, an accumulated polar feature-based DL with channel compensation mechanism is proposed to cope with the aforementioned issues. Firstly, the simulation results show that learning features from polar domain with historical information can approach optimal performance and reduce offline training overhead by 37.8 times. Secondly, the proposed neural network-based channel estimator (NN-CE) can learn the channel response and compensate for the distorted channel with 13% improvement. Moreover, in applying the novel NN-CE in the time-varying fading channel, two efficient mechanisms of online retraining are proposed, which reduce transmission overhead and retraining overhead by 90% and 76%, respectively. Finally, the performance of the proposed approach is evaluated and compared with prior arts on a public dataset to demonstrate its great efficiency.
