Implementation of a Smart Antenna System on a General-Purpose Digital Signal Processor Utilizing a Linearized CGM

The objective is to present a practical method of implementing a beamforming module with a general-purpose digital signal processor. The proposed module provides a beam-pattern that has its maximum gain along the direction of a target mobile subscriber using real-time processing. The new technique utilizes a modified CGM (conjugate gradient method) in such a way that the suboptimal gain vector of the array is obtained from the solution of an eigenvalue problem. The solution is carried out at every snapshot upon the arrival of a new signal vector. By decomposition at the autocorrelation matrix of the array inputs, the computational load of the conventional CGM, i.e., O(3N² + 12N), has been reduced significantly to the order of O(16N) in the proposed technique, where N denotes the number of antenna elements of the array. The performance of the proposed system is evaluated utilizing three different criteria; first, the execution speed for the implemented beam-former to trace the target subscriber; second, the BER (bit error rate) as a function of a given signal-to-noise ratio; and finally, the achievable capacity of the cell as a function of required BER.