The forced oscillation technique (FOT) is an active pulmonary function measurement technique that was applied to identify the mechanical properties of the respiratory system using external excitation signals. FOT commonly includes single frequency sine, pseudorandom and periodic impulse excitation signals. Aiming at preventing the time-domain amplitude overshoot that might exist in the acquisition of combined multi sinusoidal pseudorandom signals, this paper studied the phase optimization of pseudorandom signals. We tried two methods including the random phase combination and time-frequency domain swapping algorithm to solve this problem, and used the crest factor to estimate the effect of optimization. Furthermore, in order to make the pseudorandom signals met the requirement of the respiratory system identification in 4-40 Hz, we compensated the input signals' amplitudes at the low frequency band (4-18 Hz) according to the frequency-response curve of the oscillation unit. Resuts showed that time-frequency domain swapping algorithm could effectively optimize the phase combination of pseudorandom signals. Moreover, when the amplitudes at low frequencies were compensated, the expected stimulus signals which met the performance requirements were obtained eventually.
强迫振荡技术（FOT）是一种主动式肺功能测量技术，通过给呼吸系统外加激励信号的方式，辨识呼吸系统力学特性。FOT 技术常用的激励信号包括单频正弦、伪随机和周期性脉冲三种。本文针对伪随机多正弦组合信号存在的时域幅度过冲问题，研究了伪随机信号的相位优化，尝试了随机相位组合以及时-频域交换算法，以波峰因数评价优化效果。进一步根据振荡单元的频率响应曲线对低频段（4～18 Hz）输入信号的幅值进行了补偿，使最终产生的伪随机信号的时-频域特性在 4～40 Hz 范围内能够满足呼吸系统辨识要求。研究结果表明，时-频域交换算法能够有效优化伪随机信号的相位组合，振荡单元的幅频特性经过低频补偿后，能够产生满足性能要求的激励信号。.
Keywords: amplitude-frequency characteristic; forced oscillation technique; phase; pseudorandom signals; time-frequency domain swapping algorithm.