Wood Nathan A, Patronik Nicholas A, Zenati Marco A, Riviere Cameron N
Robotics Institute, Carnegie Mellon University, Pittsburgh, PA 15213, USA.
Annu Int Conf IEEE Eng Med Biol Soc. 2011;2011:7041-4. doi: 10.1109/IEMBS.2011.6091780.
HeartLander is a small mobile robot which adheres to and navigates over the surface of the heart to provide therapies in a minimally invasive manner. HeartLander's ability to efficiently operate in this dynamic environment is greatly affected by physiological motion, namely the cardiac and respiration cycles. Synchronization of robot motion with minimal intrapericardial pressure results in safer and more efficient travel. The work presented models the physiological components of motion using Fourier series and estimates their parameters using an Extended Kalman Filter. Using the Fourier series parameters, estimates of physiological phase values are calculated to be used for step synchronization. The proposed methods are demonstrated on data from a HeartLander animal study for four locations on the heart. Mean respiration phase estimates are shown to be within 5% of the true respiration phases, while mean cardiac phase estimates are shown to have a minimum error of 11%.
HeartLander是一种小型移动机器人，它能附着在心脏表面并在其表面导航，以微创方式提供治疗。HeartLander在这种动态环境中高效运行的能力受到生理运动（即心脏和呼吸周期）的极大影响。机器人运动与心包内最小压力的同步可实现更安全、更高效的移动。所展示的工作使用傅里叶级数对运动的生理成分进行建模，并使用扩展卡尔曼滤波器估计其参数。利用傅里叶级数参数，计算出生理相位值的估计值，用于步长同步。在一项针对心脏四个位置的HeartLander动物研究的数据上验证了所提出的方法。结果表明，平均呼吸相位估计值与真实呼吸相位的偏差在5%以内，而平均心脏相位估计值的最小误差为11%。
