Research On Synthesis Of Sparse Antenna Array By Alternating Convex Optimization

After nearly a century of development,modern antenna technology has been used in military,civilian and commercial areas,In modern electronic technology,in many applications such as multi-function radar and mobile communication,it is necessary and important to be able to radiate a reconfigurable multiple patterns only by changing the excitation distributions.Therefore,many algorithms have been proposed to solve this problem,such as the traditional synthesis methods including Dolph-Chebychev method,the Taylor method,the Woodward-Lawson method,the Fourier method,and Schelkunoff method,and modern synthesis methods including simulated annealing,genetic algorithm,convex optimization,particle swarm optimization,artificial neural network.Among them,the traditional synthesis methods can be very effective in designing array under some simple requirements.However,nowadays under many simultaneous complex requirements,traditional methods are generally powerless.Modern algorithms are used due to their high latitude for constraints and excellent performance.Compared with the single-beam pattern array,the multi-beam pattern antenna array can generate multiple independent high-gain directional beam patterns to jointly cover a wide space range,which is very common in multi-user mobile communication and multi-target detection systems.However,the multi-beam pattern antenna array requires multiple excitation distributions in the form of RF or digital signals,which significantly increases the costs in various aspects.In this case,reducing the number of array elements in such kind of antenna array becomes an effective way to reduce cost.Due to this,we proposes a sparse array synthesis method based on the convex optimization algorithm.This algorithm is able to give a sparse array solution under minimum distance constraint.We also give some examples to prove that this algorithm can get well-designed achievable sparse antenna arrays with scannable multiple beam constraints,side-lobe level constraints,null region constraints,and even frequency invariant constraints.All result shows the proposed method could work well and give an excellent solution.Then we propose a new antenna array structure for DOA estimation,which is a four-unit dipole ring array.The DOA performance of proposed structure is evaluated by the multiple signal classification algorithm and Cramer-Rao boundary.The results show that compared with the non-straight-connect structure with the same size,the DOA estimation error is reduced,which proves the accuracy improvement of the direction estimation.