为了提高光纤陀螺输出信号的精度和实时性,提出了一种适合在数字信号处理器(DSP)上实现的快速算法,该算法不需要传统小波变换中内插和抽取步骤,给出了相应的快速分解和快速重构实现公式,并采用半软阈值作为快速小波变换的重构阈值.实验结果表明:在静态和动态陀螺输出条件下,在TMS320C6713DSP实现半软阈值快速小波变换处理信号时间比原来降低了一个数量级,而滤波精度与原来相当,并且易于硬件实现,完全满足光纤陀螺高精度和高实时性的要求.
To improve the precision and the real-time ability of fiber optic gyroscope (FOG), a fast algorithm was developed to be uued in digital signal processor (DSP) in real-time without the processing of decimation and interpolation. Decomposition and reconstruction formula was given in detail and semi-soft threshold was selected as reconstruction threshold of fast wavelet transform. Experiment results show that time of processing the stationary and the dynamic FOG signal reduces greatly with the pro- posed algorithm realized in TMS320C6713DSP with the same precision as the traditional wavelet transform method. The semisoft threshold fast wavelet transform is able to effectively shorten the signal processing time, improve the real-time ability and furthermore it is convenient for hardware implementation with the same precision.