Comparison of phase optimization algorithms for particle manipulation via acoustic tweezer

Acoustic-based manipulation is a safe and non-damaging way to handle objects without direct physical contact, and has many potential applications in the manipulation of biological cells and micro-robots. We present a calculation model to optimize the phases of such arrays to generate different shapes of acoustic tweezers. The optimized phase combinations of transducers were determined by both genetic and Broyden–Fletcher–Goldfarb–Shanno (BFGS), and it was found that the BFGS algorithm was suited to small arrays while the genetic algorithm was faster when the array number was large. Two tweezer shapes, twin trap and vortex trap, were realized by tuning the weight parameters in the optimization model. Using the platform of a field-programmable gate array (FPGA) controller, a 1500-μm-diameter polyvinyl chloride (PVCS) bead was levitated and moved along a defined path, which verified the feasibility of the optimization model. Moreover, the control performance of the particle was quantified by image detection, showing that the moving accuracy was controlled within 583.62 µm.