Narrow Results

A novel healthy state monitor method of upper limb for space flight task is proposed. Without taking other complex diagnosis equipment in orbit, this method only uses the ordinary exercise instruments to collect and analyze the multiple muscle forces of astronauts, and deduces where the serious muscle atrophy occurs in their muscle groups of upper limb. First, the typical multiple muscle forces data of upper limb are accumulated. A 45-day 6-degree head-down tilt bed rest experiment together with a multiple muscle forces test experiment are carried out to collect the corresponding data. These data include both the muscle force data of healthy state and the related data of unhealthy state. Second, the Wavelet Packet Transform (WPT) and the Empirical Mode Decomposition (EMD) methods are used to compute the signal features of these data above. Third, a Support Vector Machine (SVM) classifier is trained by the related signal features. Finally, the trained SVM can be utilized to evaluate the healthy state of upper limb in orbit for astronaut. If the output of SVM is negative, the C-means method and the Euclidean distance can be used to locate the abnormal muscle forces and muscle groups. The concept of typical muscle group health state evaluation for upper limb is emphasized in this paper. The comparisons among the traditional diagnosis-based method, the electromyogram (EMG)-based muscle forces analysis method, and the proposed method are made. Many experiment results on ground have verified the effectiveness of proposed method.

In this paper, the energy value extracted from echo wave through wavelet transform is used as the feature of radar target, and fuzzy neural network is used as the classifier of target identification. The tested results show that the method have high classification ability.