A variable length linear array for smart antenna systems using partial optimization

Adaptive arrays have been used extensively in wireless communications applications to reduce interference between desired users and interfering signals. Significant research has been directed at the uniform linear array (ULA), however, generally a fixed array length is used. In a mobile environment this has the disadvantage of producing a fixed beamwidth due to the fixed array lengths. The variable linear array (VLA) algorithm is a proposed algorithm that allows the beamwidth to be varied by activating and deactivating elements depending on the interferers present. This approach uses the LMS algorithm for adaptation, in an embedded system this could prove costly. This paper proposes an alternative approach based on partial optimization techniques that remove the LMS from the optimization process. Furthermore studies into the cost of having such a system is presented and compared to the operational cost of the LMS algorithm.