摘要评估惯性测量单元（IMU）传感器在监测呼吸周期的准确性和可行性。通过IMU传感器对20名健康志愿者进行了5分钟的腹壁运动监测，经过滤波处理的加速度数据转化为波形图，从中提取波峰和波谷以计算呼吸周期。理想轴与对照轴的数据进行比较及其误差分析、相关性分析、线性回归分析、Bland-Altman分析。研究结果显示，本研究所采用的算法能够在1698个呼吸周期样本中准确识别呼吸周期，平均绝对误差为0.14 s/呼吸周期，相关系数高于0.86，Bland-Altman分析的平均差异为0.01，证明了IMU传感器在监测呼吸周期方面的高度可靠性，为开发成本效益高、便携、实时且稳定的呼吸监测系统提供了新的研究方向和方法论。

abstractsThis study aimed to evaluate the accuracy and feasibility of Inertial Measurement Unit (IMU) sensors in monitoring respiratory cycles. The IMU sensors were used to monitor the abdominal wall movement of 20 healthy volunteers for 5 minutes. The collected acceleration data were processed through filtering algorithms and transformed into waveform graphs, from which peaks and troughs were extracted to calculate the respiratory cycles. The data from the ideal and control axes were compared, and their errors were analyzed, along with correlation analysis, linear regression analysis, and Bland-Altman analysis. The results showed that the algorithm used in this study could accurately identify respiratory cycles in 1698 samples, with an average absolute error of 0.14 seconds per respiratory cycle, a correlation coefficient greater than 0.86, and an average difference of 0.01 in Bland-Altman analysis. This demonstrates the high reliability of IMU sensors in monitoring respiratory cycles and provides a new direction and methodology for developing cost-effective, portable, real-time, and stable respiratory monitoring systems.