Study On Antenna Array Pattern Novel Fast Synthesis Method

Nowadays, antenna arrays have been widely utilized in radar and communication applications. Desired pattern can be achieved by adjusting the parameters such as the element number, arrangement, spacing, location, excitation distribution and so on, that is hard to be met for an antenna unit. On the contrary, the process of determining the parameters to seek the desired pattern is known as the antenna array pattern synthesis.There are several classic methods for antenna array pattern synthesis, such as Dolph-Chebyshef method, Taylor method, perturbation method, Woodward-Lawson sample method, Fourier series method and so on. Good results may be obtained by these classic methods for specific problems, but it is always helpless for complex problems, for example, the multi-objective and multi-parameter optimization problems. The dissertation will focus on the following four aspects to study the problems of antenna array pattern synthesis:firstly, find a generalized, effective synthesis method to deal with sorts of optimization problems; secondly, the coupling effect and the actual excitation distribution should be considered to exactly design the antenna systems; thirdly, the problems of occupied memory and time-cost for large-scale optimization problems; fourthly, fast method for calculating the antenna array patterns.In this dissertation, researches have been done on these problems and the main contributions are as follows:1. Inspired by the dictatorial system, we present the improved multi-population leader dominating genetic algorithm (MPLDGA) by combining the dictatorial system and traditional genetic algorithm The algorithm is suitable for handling optimization problems with merits of fast convergence rate, robust global searching ability and less time-cost.The patterns and side-lobes are the most common optimization objectives. Method is presented to deal with the kind of problems by analyzing the relationship between the patterns and side-lobes. The patterns are regarded as the primary goal and optimized firstly, and then, the side-lobes are optimized with the optimized patterns altering slightly. The proposed MPLDGA is adopted to deal with both continuous and combinational optimization problems and satisfied results are reached.2. Microwave cascade networks have been studied. The ideal lossless symmetrical reciprocal network (ILSRN) is constructed and introduced to resolve the complex interconnections of two arbitary microwave networks. By inserting the ILSRNs, the complex interconnections are converted into the standard one-by-one case without changing the characteristics of the previous microwave networks. Based on the algorithm of the generalized cascade scattering matrix, a useful derivation on the excitation coefficients of antenna arrays is firstly proposed with consideration of the coupling effects and we present a fast precise method for the antenna array pattern synthesis.The proposed technique is applied both on the microwave circuits and antenna arrays. In the aspect of microwave circuit, a wideband high-isolation magic-T is optimized. Its equivalent circuit is segmented into sub-networks, which are calculated by the proposed technique and embodied in the genetic algorithm. The optimized magic-T provides merits of good impedance matching, high isolation, good imbalance and wideband. In the aspect of antenna arrays, A novel fast method for designing slotted waveguide antennas is presented, and the traditional Elliott’s method is replaced by our method. According to the method, two kinds of slotted waveguide arrays are successful designed. Finally, advanced modifications are adopted to rectify the undesired performances by exactly predicting the performances and modifying the feeding network of a two-unit antenna array.3. A novel partitioned strategy is presented to deal with the large-scale optimization problems. A large number of optimized parameters are stochastically partitioned into a set of blocks, which will be optimized one by one by any suitable algorithm with the other optimized parameters contained by the other blocks as a feedback. The occupied memories and the time-cost are reduced greatly, and the traditional optimization algorithms (for example, tht genetic algorithm) can be adopted to handle the large-scale optimization problems. The optimized parameters, which will take full advantage of the merits of evolutionary algorithms, are initialized with stochastic values and no prior knowledge is required. In this dissertation, the partitioned strategy and hybrid genetic algorithm are adopted and two antenna arrays are thus optimized designed. Firstly, a contoured beam covering the mainland of the P.R. China is achieved for a 3,600-element antenna array. Secondly, the phases of a 10,000-element antenna array are optimized to minimize the maximum side-lobes with given gain losses.4. The GPU (Graphic Processing Unit) technology is adopted for antenna array pattern synthesis, and the parallel computation is used to accelerate the time-cost of antenna array pattern synthesis with the consumed time reduced greatly. In this dissertation, the traditional serial computation is transformed into the parallel computation, and both the serial (based on CPU) and parallel (based on GPU) double chains quantum genetic algorithm are used to optimize the side-lobes of sparse arrays. The results show that a 58X speedup ratio is achieved and the side-lobes are reduced.