The limited quantization digit of voltage and the effect of fringing field between adjacent electrodes (phased-array
controlling units) limit the deflection efficiency when a Liquid-Crystal Phased Array (LCPA) is used for beam steering.
In this paper, an optimization algorithm named pattern search is proposed to improve the diffraction efficiency. This
algorithm directly optimizes the step phase slope to obtain high diffraction efficiency, rather than discussing the complex
relationship between the diffraction efficiency and various influence factors. Besides, other optimization algorithms
based on phase retrieval, such as GS, need the entire energy distribution; however, it is hard to obtain in practical. Our
algorithm need only the energy of the target diffraction point and it can be easily realized. Firstly, we construct the model
for beam steering, and point out that the deformation of the phase slope by influence factors is the reason why
conventional method can not realize high diffraction efficiency. Secondly, we construct an optimization model for the
issues and apply the pattern search algorithm to optimize the diffraction efficiency. The simulation results show high
performance of our algorithm comparing with the conventional steering method. Finally, a set of beam steering
experiments were performed with a one-dimensional LCPA being set both according to the un-optimized and the
optimized recipe, and the results were in very good agreement with the theoretical predictions. We show that the
deflection efficiency can be drastically improved.
In this paper, by establishing a liquid crystal model of beam steering, a rapid iterative algorithm called Rosenbrock
algorithm is proposed for obtaining wave-control data of liquid crystal phased array (LCPA). The method doesn’t need
wavefront reconstruction, the error is reduced, and the searching direction is adjusted automatically to right space of
solution, thus the algorithm converge rapidly, and the performance of beam steering is optimized at the same time. In this
paper, firstly, a liquid crystal model of beam steering is established. Then, the effectiveness and superiority of this
algorithm is verified compared with Pattern Search algorithm. Finally, we evaluate our algorithm by experiment. The
results show that Rosenbrok algorithm can optimize the efficiency of beam steering rapidly and significantly.