Pattern Synthesis Of Antenna Array Using Genetic Algorithm

With the development of wireless communication, the limited frequency resource declined rapidly. To optimize the resource allocation and improve the utilization ratio of resource is becoming a pursuing goal for scientists and engineers. Smart antenna technology occurred in such background, and pattern synthesis of antenna array is the core technology of smart antenna. It can utilize the resource effectively by adjusting the parameters of the antenna units such as the numbers, positions, feeding-back amplitudes and phases in different frequency circumstance. The comparative advantage of genetic algorithm is obvious because of its' ability of global rapid optimum.Genetic algorithm is a bionic algorithm which simulates the heredity and variation of creature. Based on the principle of "the survival of the fittest", it can find the optimal solution through the process of selecting, crossing and variation. Genetic algorithm has the advantages to settle large space, multi-parameter and no-linear problems, so it is applied in extensive field. In recent years genetic algorithm has application to analyze and design antenna.Mathematical model of array signal processing was introduced in this paper. Decimal coded genetic algorithm was theorized by analyzing on the equivalency between decimal and binary coded GA. While controlling the narrow-range nulls in the pattern synthesis, several controlling factors affecting iterative efficiency were investigated and structure of genetic algorithm was improved. And the simulation proved the algorithm was effective. In order to solve the problem of controlling the broad-range nulls in pattern synthesis of antenna array, the process of selecting elitist was chosen in the algorithm. Selecting and crossing operation was developed. To get more ideal chromosome, the complex-coded genetic algorithm based on sorting was further improved. In the improved algorithm the crossing and selecting probability was adopted, and the crossing operation included several processes such as interpolation, extrapolation and so on. The variation operation was also developed. Successful simulation shows that genetic algorithm has better convergence and practicability.