| Heart sound signal is a symbol signal reflects the human’s heart is healthy or not, it contains a lot of useful information reveals the state of heart and its pathological conditions. In recent decades, people have been trying to simplify and abstract this complex physiological system as an ideal linear model. By the analysis of heart sound and ECG signal in traditional time domain, frequency domain and time-frequency combination methods, a series of remarkable achievements have been made.However, the essence of heart is a complex nonlinear dynamical system, methods of traditional linear analysis are not sufficient to study the nonlinear activities of life. Chaotic motion is an extremely important movement in nonlinear system, analysis and research based on chaos theory can reveal the inherent special laws within the randomness of nonlinear system. Therefore, this paper intends to study the heart sound signal from this point of view, it will deeply explore and reveal the inherent dynamic characteristics of non-stationary heart sound signal in nature. The main research work is as follows:1. Give a nonlinear model for the heart sound signal, a chaotic system similar with heart sound signal waveform is proposed, then the paper calculates the corresponding chaotic characteristics of the system, and gives the time-domain waveform and attractor phase diagram. Finally, using the way of similar factor combined with similar phase diagram proves that time-domain waveform generated by the system has a certain similarity with the real heart sound signal. This provides an idea and method for the nonlinear analysis of heart sound signal.2. Using the shoulder belt heart sounds acquisition device designed by our team to collect different motion state and different age group of heart sound signals, and calculate the appropriate delay time with embedding dimension in order to reconstruct the phase space of heart sound signals, and then begin to analyze the chaotic characteristics. Conclusion(1): heart sound correlation dimension is bigger at first, then it will reduce during exercise, and finally recover after the exercise. Conclusion(2): With people age increasing, the heart sound signal correlation dimension and Lyapunov index showed a downward trend. These conclusions are consistent with the relevant ECG chaotic characteristic findings. Combined with the clinical diagnosis, these indexes can be more in-depth study of the variation of cardiac status.3.In the field of signal processing, building time series forecasting model is an important and innovative research direction, this paper introduces the Volterra prediction method based on the theory of chaotic time series and uses it in the application field of heart sound signal short-term prediction, this research can help the user know the future changes of heart sound in 3 ~ 5s, which facilitates doctor’s early intervention, besides can be used for surgery, tooth extraction and other practical applications. At the same time, this paper proposes a heart sound long-term forecasting model based on empirical formula, the long future changes of heart sound waveform can be given an as an estimation, which facilitates the user to understand the development trend of his or her heart sound.Finally, in order to meet the requirement of engineering application, and help the users detect their heart sound chaotic index and predict the waveform in advance, the paper puts forward an application platform gathering chaotic analysis and forecasting capabilities designed by Matlab GUI. |
PDF Full Text Request