摘要
提出一种基于时变多变量回归模型利用ECG信号和心脏运动信号相关性进行预测的预测方法,该方法将ECG信号非平稳心率变化信息的全过程通过模型中互相关项引入心脏信号预测中,以克服单纯的基于线性模型预测方法的不足,增强了对心脏信号中非平稳变化的适应能力,提高了估计的精度。通过对偶卡尔曼滤波器对模型的状态和参数分别进行估计,完成心脏运动信号的实时预测,并通过比较实验验证了算法的有效性。
The high bandwidth and high amplitude features of the beating heart motion make it difficult for the doctor to operate in the off-pump coronary artery graft bypass surgery. The precise tracking control algorithm is the core of the robotic assisted system and the accurate beating heart motion prediction is the core of robot tracking control procedure. A prediction method is proposed to take into account the small changes of heart rate and interaction nature between ECG signal and heart motion signals to improve the heart motion prediction. This algorithm is based on the heart motion mathematical representa- tion by using an adaptive time varying multivariate vector autoregressive (MVAR) model. The model is parameterized by dual Kalman Filters to estimate its states and parameters respectively. The comparative experiments results for evaluating the proposed algorithm are reported by using the vivo collected data.
出处
《天津职业技术师范大学学报》
2015年第1期1-5,共5页
Journal of Tianjin University of Technology and Education
基金
国家自然科学基金资助项目(61178048)
天津市应用基础与前沿技术研究计划(14JCQNJC04300)
天津职业技术师范大学科研发展基金资助项目(KJY11-10
KYQD13022)