摘要

This paper presents a beam pattern synthesis method for low cost implementation of antenna arrays. Instead of designing array weights, our method synthesizes unequally-spaced array with uniform antenna weights. The effect of mutual coupling is considered when we search a set of optimum antenna positions. The determination of the antenna positions is a nonlinear optimization problem because the positions are arranged as the parameters of an exponential or trigonometric function. This nonlinear optimization problem is solved by using an improved genetic algorithm. Since the process of digitization and complex weighting (amplification and phase-shifting) is not required in the implementation of the designed array, the complexity and cost of antenna array implementation can be significantly reduced. Experiment results also show that designing the positions of antennas in an array is more likely to achieve lower sidelobe level than designing the array weights. These properties support that the proposed method is suitable for practical applications.