Lead-free potassium sodium niobate piezoelectric single crystals substituted with lithium 0.95(K0.5Na0.5)NbO3–0.05LiNbO3 have been grown by Bridgman method and their dielectric and piezoelectric properties were studied. The orthorhombic-tetragonal and tetragonal-cubic phase transition temperatures of the single crystal appear at 192 and 426°C according to the dielectric constant versus temperature loops, respectively, and the (001) plates show good piezoelectric properties with piezoelectric constant d33 as high as 405pC∕N, large thickness electromechanical coupling factor kt=61%, and low dielectric constant of 185 at room temperature. These excellent properties show that the 0.95(K0.5Na0.5)NbO3–0.05LiNbO3 single crystal is a good lead-free piezoelectric material.

Relaxor-based ferroelectric single crystals Pb(In1∕2Nb1∕2)O3–Pb(Mg1∕3Nb2∕3)O3–PbTiO3 (PIMNT) have been grown directly from their melt using the vertical Bridgman method, and their boules have reached the size of ϕ45×80mm. The as-grown PIMNT28/40/32 crystals on the (001) cuts exhibit a dielectric constant ε∼5200, dielectric loss tanδ∼0.50%, piezoelectric strain constant d33∼1700–2200pC∕N, electromechanical coupling factors kt∼0.61 and k33∼0.92, coercive field Ec∼10.88kV∕cm, remanent polarization Pr∼46μC∕cm2, Curie temperature TC∼192°C, and rhombohedral to tetragonal phase transition temperature Trt∼119°C. Moreover, their piezoelectric properties show good thermal stability under the heat treatment at 105°C.

The Convolutional Neural Network (CNN) has been used in many fields and has achieved remarkable results, such as image classification, face detection, and speech recognition. Compared to GPU (graphics processing unit) and ASIC, a FPGA (field programmable gate array)-based CNN accelerator has great advantages due to its low power consumption and reconfigurable property. However, FPGA’s extremely limited resources and CNN’s huge amount of parameters and computational complexity pose great challenges to the design. Based on the ZYNQ heterogeneous platform and the coordination of resource and bandwidth issues with the roofline model, the CNN accelerator we designed can accelerate both standard convolution and depthwise separable convolution with a high hardware resource rate. The accelerator can handle network layers of different scales through parameter configuration and maximizes bandwidth and achieves full pipelined by using a data stream interface and ping-pong on-chip cache. The experimental results show that the accelerator designed in this paper can achieve 17.11GOPS for 32bit floating point when it can also accelerate depthwise separable convolution, which has obvious advantages compared with other designs.