solutions that provide seamless communication for mobile users. However, with the development of various wireless networks, the spectrum detection of cognitive networks and terminals becomes more complicated, which decreases the detection performance. The accuracy and efficiency of the spectrum detection will be reduced due to integrating various wireless networks with different characteristics into a diverse overlay system. In this paper, we design a high accuracy recognition algorithm for Cognitive Radio (CR) signal based on machine learning in HWNs, which can recognize the received signal type through extracting the features. This algorithm can recognize the signal types blindly with low complexity, and prevent the influence of “hostile terminals”. Simulation results indicate that the algorithm we proposed can achieve high recognition accuracy under either Additive White Gaussian Noise (AWGN) channel or Rayleigh fading channel.

Index Terms—Signal recognition, cognitive radio, machine learning, SVM, heterogeneous wireless networks

Cite: Jian Liu, Jibin Wang, and Sani Umar Abdullahi, "High Accuracy Signal Recognition Algorithm Based on Machine Learning for Heterogeneous Cognitive Wireless Networks," Journal of Communications, vol. 12, no. 3, pp. 173-179, 2017. Doi: 10.12720/jcm.12.3.173-179