Efficient Beamformer for Low Mobility Indoor Communication Using Sparse Adaptive Algorithm

The proposed beamforming model exploits the underlying sparseness of the adaptive filter as impulse response of wireless channel shows some extent of sparse behavior in practice. A new formulation of cost function of fourth order instead of quadratic will help to achieve stable and faster convergence, but the computational complexity is high. Thus the filter design requires a compromise between the quadratic and fourth order cost function to achieve good estimation accuracy. Normalized least mean square fourth (NLMS/F) filter design is based on the compromise to achieve a better performance with a faster convergence. Inclusion of sparsity in the cost function of NLMS/F filter further reduces the computational complexity as less number of nonzero coefficients involve in estimation with a bounded error. IEEE 802.11 and Saleh–Valenzuela models with exponential power delay profile (PDP) are used to implement proposed beamformer for indoor application. The proposed sparse-NLMS/F beamformer is compared with its NLMS/F counterpart, and tested with practical fading condition. Different performance measures like mean square error (MSE), estimated weight convergence, beam pattern are used to test the proposed model under practical channel condition.