Improved Homptopy Technique for MIMO signal Detection

In this paper, an improved Homotopy algorithm has been presented for signal detection in a MIMO system. The channel matrix is trained as a learnable matrix which results in better signal update. Furthermore, a smooth map is adopted to increase the signal detection performance. The nonlinear denoiser is used to remove the noise more efficiently. Also, the hyperbolic cosine loss function has been adopted to smoothen the results. The simulations show that the proposed method is superior to its counterparts in the bit error rate performance.