Abstract:Some major global optimization algorithms, which involves all the optimized variables in one problem, were used to synthesize time-modulated linear antenna arrays (TMLAs) in the existing literature. In this work, a hybrid approach, combining differential evolution (DE) algorithm and convex optimization, is proposed. This hybrid method is performed to optimize the static excitations amplitude and the switch-on duration time of TMLAs in order to achieve ultralow sidelobe and low sideband radiation at the same time effectively. In one iteration, as long as a set of switch-on duration time is fixed by the DE algorithm, Then the problem with respect to the static excitation is convex. So a convex optimization programming can solve the static excitation which is best for the corresponding set of switch-on time. The approach makes full use of the convexity of the problem with respect to the static excitation variables. Consequently, DE algorithm is used to optimize the switch-on time variables only. Which means fewer optimized variables are needed and it is more likely to find the global optimal solution in a less iterations. So that faster convergence can be obtained compared to DE algorithm which includes all optimized variables, especially for larger arrays. Furthermore, some examples are presented to demonstrate the effectiveness of this hybrid approach. First is a 32-element broadside TMLAs, with half wavelength uniform spacing, which is concerned for synthesizing a ultralow sideband at the same time suppress the sideband radiation pattern. Second is to sparse a uniform spaced 30-element TMLAs with fewer element satisfies the requirement of pattern. The last is to synthesis shaped beams. For the first example, numerical results show that a -61.4 dB sidelobe pattern with a static excitation amplitude ratio 4 can be synthesized, while the sideband level (SBL) is suppressed to -32.1 dB. For the second example, we can only use 18-element to achieve a -23.8 dB sidelobe and -31.3 dB SBL pattern with a static excitation amplitude ratio 2.77.

Applying Automatic Differentiation to Optimize Differential Microphone Array Designs

Equivalent linear analysis and optimization of tracking differentiator

Application of Convex Optimization to Acoustical Array Signal Processing

Theoretical Lower Bounds on the Performance of the First-Order Differential Microphone Arrays With Sensor Imperfections

First-order Differential Microphone Array Based on FLMS Adaptive Speech Enhancement Method

Design of Differential Loudspeaker Line Array for Steerable Frequency-Invariant Beamforming

Design of a Wideband Linear Microphone Array for High-Quality Audio Recording

An Improved Genetic Algorithm for Thinning Acoustic Array.

First-Order Differential Microphone Array For Robust Speech Enhancement

On the Design of Robust Differential Beamformers From the Beampattern Error Perspective

Study of Speech Enhancement Based on the Second-Order Differential Microphone Array

A scalable hybrid analog-digital architecture for multi-channel feedforward active noise control

An Effective Hybrid Optimization Algorithm for the Synthesis of 4-D Linear Antenna Arrays

DHASP: Differentiable Hearing Aid Speech Processing

Constant Directivity Loudspeaker Beamforming

Thinning and Weighting of Planar/Conformal Arrays Considering Mutual Coupling Effects

A Momentum Two-gradient Direction Algorithm with Variable Step Size Applied to Solve Practical Output Constraint Issue for Active Noise Control

Low Complexity Design of Logistic Distance Metric Adaptive Filter for Impulsive Noise Environments

An iterative LMI approach to IIR noise transfer function optimization for delta-sigma modulators

Implementation of Two-Poly Differential MEMS Microphones for SNR and Sensing Range Enhancement

Liapunov Analysis and Design of Linearly Constrained Adaptive Filtering Algorithms