【作者】 王玉峰；

【导师】 庞伟正；

【作者基本信息】 哈尔滨工程大学 ， 电磁场与微波技术， 2007， 硕士

【摘要】 阵列天线广泛应用与雷达、无线通信和电子对抗等领域，方向图综合作为阵列天线的关键技术，在抗干扰、截获率和参数估计等方面有着重要的作用。阵列天线方向图函数的最优化是一种非线性优化问题。近年来，遗传算法由于其优异的全局搜索性能且对目标函数无可微和连续性等条件的限制，广泛应用于阵列天线综合领域。免疫算法是近年来出现的一种与遗传算法相似，基于智能搜索的全局优化技术，有学者研究表明免疫算法有着比遗传算法更强的函数优化性能。本文对免疫算法在阵列天线方向图综合中的应用方面进行了研究。对旁瓣抑制和零点生成进行了讨论，并对其中的超低副瓣天线阵和稀疏阵列的低副瓣控制作了详细的研究；结合免疫算法和混沌优化各自优点，提出了多映射混沌免疫算法并将其应用到阵列天线综合中，取得了很好的优化效果。本文在理想条件下，运用免疫算法对直线阵和矩形阵进行了方向图综合的仿真，得到如下结论：1．免疫算法是阵列天线方向图综合的有效方法。免疫算法是一种启发式的高效智能仿生算法，它采用分散和独立搜索方式保证群体的多样性，从而避免了遗传算法单一搜索模式带来的早熟现象。文中采用免疫算法分别对直线阵和矩形阵的阵元间距、激励幅度和相位进行了优化，并根据指标要求实现了对副瓣电平的有效控制、超低副瓣的生成及在指定角度生成指定零深的零陷。2．免疫算法对方向图综合效果优于遗传算法，并且处理的优化问题越复杂，效果越明显。本文为了更客观的研究免疫算法应用于方向图综合的效果，对同一问题还采用遗传算法进行了仿真。结果表明，采用免疫算法综合后的方向图明显优于采用遗传算法的，并且当阵元的数目越多或方向图函数的形式越复杂，免疫算法的优越程度表现的就越明显。3．采用了离散化阵列天线方向图函数和克罗内克积形式的矩形稀疏阵列方向图函数。对阵列天线方向图函数进行采样，采用适合计算机和免疫算法运算的离散化函数，用矩阵运算替代了循环迭代，提高了计算的效率，并在仿真中得到验证；将矩形稀疏阵方向图函数改写成只有两个变量的克罗内克积形式，特别适合Matlab运算，同时也提高了算法的搜索效率。4．融合混沌优化和免疫算法各自的优势，提出了多映射混沌免疫算法。用混沌序列的遍历性和混沌扰动有“规律”的随机性对免疫算法的全局搜索性能进行进一步的优化，使用了两种不同规则的混沌序列完成免疫算法中不同阶段的抗体生成，保证抗体群具有足够的多样性，使得搜索范围足够大，克服了早熟现象；使用了混沌扰动与免疫算法自身变异操作相结合的混合变异方法，提高算法的搜索效率和收敛速度。将多映射混沌免疫算法应用到阵列天线方向图综合中，验证了算法搜索能力强，能有效避免局部收敛，达到全局优化。更多还原

【Abstract】 The applications of antenna arrays can be found in many areas such as radar, wireless communications and electronic warfare. As the key technologies in antenna array, pattern synthesis have been widely used in interference suppression、probability of intercept and parameter estimation.The optimization of pattern function is non-linear problem. In resent years, As its global search performance and having no restraint of differentiable and continuities, Genetic algorithm has been widely used in pattern synthesis of antenna array. Immune algorithm which is a new global optimal technology based on intelligent search is very similar to genetic algorithm. Some scholars indicate that immune algorithm has the better performance in function optimization.This thesis carried on a research in applying immune algorithm to pattern synthesis of antenna array. Sidelobe suppression and null steering are researched. The control of low sidelob in ultra-low sidelobe antenna array and thinned array are discussed detailedly. Through the study of immune algorithm and chaos optimal algorithm, Multi-map chaos immune algorithm has been presented. The result which gained by applying Multi-map chaos immune algorithm to the synthesis of antenna array is excellent.Pattern synthesis of linear array and rectangular array using immune algorithm is under the ideal condition in the thesis. Conclusions gained are as follows:Firstly, Immune algorithm is the efficient method of pattern synthesis of antenna array.Immune algorithm is a efficient heuristic intelligent bionic algorithm. It uses dispersed and separate searching method to ensure the diversity of the colony, in order to avoid the earliness which can be produced by single searching mode in genetic algorithm. In the thesis, array spacing、amplitude and phase of excitement in linear array and rectangular array are optimized using immune algorithm. The result confirm with the design beacon which is the efficient control of sidelobe、ultra-low sidelobe and depth of the null steering in specified angle.Secondly, The performance of pattern synthesis using immune algorithm is better than which using genetic algorithm, and which is more evidence when problem required optimization is more complexity.The same problem is simulated by both immune algorithm and genetic algorithm for making the research of pattern synthesis applying immune algorithm more objective. Results show that pattern synthesized using immune algorithm is more excellent than which of genetic algorithm, and when the array amount is larger or the pattern function is more complicated, the superiority is more obvious.Thirdly, The pattern function of antenna array is transformed to discrete form, and which of rectangular thinned array is turned to the form of Kronecker product.Through sampling to pattern function of antenna array, a discrete function is gain which is adapted to computer operating and immune algorithm. Circulatory operation is replaced by matrixes’ product, which advances efficiency of computing. Pattern function of rectangular thinned array has been transformed to the form of Kronecker product by two matrixes, which not only fits the operation of Matlab ,but also enhances the searching efficiency.Fourthly, Combining respective advantages of chaos optimal algorithm and immune algorithm, Multi-map chaos immune algorithm is presented.There is a farther optimization to the global searching performance of immune algorithm using the ergodicity of chaos sequence and the randomicity with "rules" of chaos disturbance. Two chaos sequences with different rules are applied to control the generation of the antibodies in different periods, which ensure the diversity of antibodies. As a result the searching area is adequacy, earliness phenomenon is conquered. A hybrid mutation method is presented which combining Chaos disturbance and the own mutation method of immune algorithm. As a result the searching efficiency is advanced and the speed of convergence is improved. The synthesizing of antenna array using Multi-map chaos immune algorithm have gain a fine effect. The result validate that the presented algorithm has a strong searching ability and can get the global optimal value while avoiding local convergence.更多还原