Objective  In recent years, terahertz radar has been widely used in the fields such as human security and non-destructive testing due to its advantages of high resolution, high penetration, and high safety. Researchers have proposed terahertz MIMO array radar from the perspective of improving imaging speed. This radar combines the spatial division multiplexing technology of MIMO arrays to achieve fast and real-time imaging. However, due to the short wavelength of terahertz waves and the sparse design of MIMO arrays, the array element spacing is too large, resulting in high gate sidelobe levels in the radar beam pattern, which affects imaging quality. Optimizing the array position through optimization algorithms can effectively solve this problem, but previous research is mainly focused on optimizing low-frequency MIMO array radars, while high-frequency terahertz MIMO array radars may encounter more severe high gate sidelobe level problems. Therefore, it is necessary to design optimization algorithms with higher optimization accuracy for this band. Therefore, from the perspective of solving this problem, this paper first abstracts the optimization model of terahertz MIMO linear array, and then proposes a dual strategy adaptive differential evolution for array optimization based on the optimization characteristics of the model.

  Methods  A multi-constraint optimization model is established with the goal of reducing the peak sidelobe level ratio based on the optimization characteristics of terahertz MIMO arrays. Using Kent chaotic sequences to generate an initial population, this method can make the distribution of initial individual genes more uniform in the solution space. A dual mutation strategy was proposed to enable the algorithm to select appropriate mutation strategies based on the number of iterations and individual fitness values. Adaptive improvements have been made to the parameters, allowing them to be autonomously adjusted based on individual evolution. The convergence performance of the algorithm was tested through standard functions, and the effectiveness of the algorithm for terahertz MIMO array optimization was tested through simulation experiments.

  Results and Discussions   The DSADE algorithm proposed in this paper has the best optimization effect on the 8 transmitting and 8 receiving terahertz MIMO array antenna, and the optimized minimum peak sidelobe level ratio is 1.32 dB lower than the ISMADE algorithm (Tab.5). It can be clearly seen from Fig.4 that the DSADE algorithm effectively suppresses the gate sidelobe level in the directional synthesis map of the MIMO array. The comparison of the 50 optimization results (Fig.6) also shows that the optimization performance of the DSADE algorithm is significantly better than other algorithms. It has been proven that this algorithm can effectively optimize terahertz MIMO arrays, suppress gate sidelobe levels, and improve imaging quality.

  Conclusions  A portable infrared target simulation system is designed with working wavelengths of 3-5 μm and 8-14 μm. This system has the characteristics of simple structure, adjustable wavelength, rich targets, and clear and stable imaging. The wavefront quality of the system was analyzed using Zemax software, and at 4 μ the PV value of the center field of view in the m-band is 0.0132λ. The root mean square value is 0.0038λ, at 12 μ the photovoltaic value of the center field of view in the m-band is 0.0044λ. The mean square difference is 0.0013λ. An optical mechanical thermal analysis was conducted on the collimation system, and at a temperature difference of 30 ℃, the deformation caused by the mechanical structure was much greater than that of the primary and secondary mirrors themselves, reaching 10% μ. In the order of m, the imaging results have significant defocusing errors, which can be compensated for by temperature changes through refocusing the target disk in an adjustable three-dimensional position. The imaging function of the system was tested. For targets of different shapes, the system can generate clear and recognizable images, providing stable simulated targets for infrared detection equipment.