人体软组织的含水量高达65%，所以超声波在人体软组织中传播的衰减小、致病风险低且干扰范围有限，在体内通信具有优势。然而，超声人体信道具有密集多径特性，经过反射和折射多条路径到达接收端的信号交叠在一起会产生多径干扰，影响通信可靠性。现有研究通过发射极短脉冲的超声波来避免多径交叠及其带来的干扰，但大方向角的超宽带超声探头难以实现，实际探头产生的脉冲宽度不够小，在接收端会产生多径交叠，对信号判决产生的干扰不能忽略。为了研究体内多径干扰的分布特性及其对宽带超声人体通信的影响，文中利用k-Wave仿真工具包对超声人体信道进行三维建模，然后通过仿真实验获得信道冲激响应，分析多径时延分布的统计特性并进行曲线拟合，根据接收端的判决机制估算人体信道的多径干扰，推导了直接序列扩频超声波宽带体内通信的误比特率下界，并通过蒙特卡罗实验验证了这个误比特率下界的有效性。实验结果表明，信噪比较低时多径干扰不能忽略，调整扩频码长可改善通信性能。

The water content of human soft tissues is as high as 65%, so ultrasound has advantages in low attenuation, low risk of disease, and limited interference range in human soft tissues, making it advantageous for communication in the body. However, the ultrasonic human channel has the characteristics of dense multipath, so the signals that reach the receiver through multiple paths of reflection and refraction overlap to generate multipath interference, which affects communication reliability. The existing research avoids multipath overlap and its interference by emitting extremely short pulses of ultrasound, but it is difficult to achieve ultra wideband ultrasound probes with large directional angles. The actual pulse width generated by the probe is not small enough, so multipath overlap will occur at the receiver and the interference cannot be ignored in signal judgment. In order to study the distribution characteristics of multipath interference in the body and its impact on wideband ultrasound human body communication, this paper used the k-Wave simulation toolkit to model the ultrasound human body channel in 3D. Then the channel impulse response was obtained through simulation experiments, and the statistical characteristics of multipath delay distribution were analyzed and curve fitting was performed. Based on the decision mechanism of the receiver, the multipath interference of the human channel was estimated, and the lower bound of the bit error rate (BER) for direct sequence spread spectrum ultrasonic broadband (DS-UsWB) was derived. The effectiveness of this lower bound of BER was verified through Monte Carlo experiments. Experimental results indicate that multipath interference cannot be ignored when the signal-to-noise ratio is low, and the communication performance can be improved by adjusting spread spectrum code length.