Multimodal Sensor Motion Intention Recognition Based on Three-Dimensional Convolutional Neural Network Algorithm

With the development of microelectronic technology and computer systems, the research of motion intention recognition based on multimodal sensors has attracted the attention of the academic community. Deep learning and other nonlinear neural network models have a wide range of applications in big data sets. We propose a motion intention recognition algorithm based on multimodal long-term and short-term spatiotemporal feature fusion. We divide the target data into multiple segments and use a three-dimensional convolutional neural network to extract the short-term spatiotemporal features. The three types of features of the same segment are fused together and input into the LSTM network for time-series modeling to further fuse the features to obtain multimodal long-term spatiotemporal features with higher discrimination. According to the lower limb movement pattern recognition model, the minimum number of muscles and EMG signal characteristics required to accurately recognize the movement state of the lower limbs are determined. This minimizes the redundant calculation cost of the model and ensures the real-time output of the system results.

Classifying Motion Intention of Step Length and Synchronous Walking Speed by Functional Near-Infrared Spectroscopy

In some patients who have suffered an amputation or spinal cord injury, walking ability may be degraded or deteriorated. Helping these patients walk independently on their own initiative is of great significance. This paper proposes a method to identify subjects’ motion intention under different levels of step length and synchronous walking speed by using functional near-infrared spectroscopy technology. Thirty-one healthy subjects were recruited to walk under six given sets of gait parameters (small step with low/midspeed, midstep with low/mid/high speed, and large step with midspeed). The channels were subdivided into more regions. More frequency bands (6 subbands on average in the range of 0-0.18 Hz) were decomposed by applying the wavelet packet method. Further, a genetic algorithm and a library for support vector machine algorithm were applied for selecting typical feature vectors, which were represented by important regions with partial important channels mentioned above. The walking speed recognition rate was 71.21% in different step length states, and the step length recognition rate was 71.21% in different walking speed states. This study explores the method of identifying motion intention in two-dimensional multivariate states. It lays the foundation for controlling walking-assistance equipment adaptively based on cerebral hemoglobin information.

Online Adaptive Prediction of Human Motion Intention Based on sEMG

Accurate and reliable motion intention perception and prediction are keys to the exoskeleton control system. In this paper, a motion intention prediction algorithm based on sEMG signal is proposed to predict joint angle and heel strike time in advance. To ensure the accuracy and reliability of the prediction algorithm, the proposed method designs the sEMG feature extraction network and the online adaptation network. The feature extraction utilizes the convolution autoencoder network combined with muscle synergy characteristics to get the high-compression sEMG feature to aid motion prediction. The adaptation network ensures the proposed prediction method can still maintain a certain prediction accuracy even the sEMG signals distribution changes by adjusting some parameters of the feature extraction network and the prediction network online. Ten subjects were recruited to collect surface EMG data from nine muscles on the treadmill. The proposed prediction algorithm can predict the knee angle 101.25 ms in advance with 2.36 degrees accuracy. The proposed prediction algorithm also can predict the occurrence time of initial contact 236±9 ms in advance. Meanwhile, the proposed feature extraction method can achieve 90.71±3.42% accuracy of sEMG reconstruction and can guarantee 73.70±5.01% accuracy even when the distribution of sEMG is changed without any adjustment. The online adaptation network enhances the accuracy of sEMG reconstruction of CAE to 87.65±3.83% and decreases the angle prediction error from 4.03∘ to 2.36∘. The proposed method achieves effective motion prediction in advance and alleviates the influence caused by the non-stationary of sEMG.