A Novel Estimation Approach of sEMG-based Joint Movements via RBF Neural Network

Intention Recognition of Elbow Joint based on sEMG Using Adaptive Fuzzy Neural Network

“…In this subsection, the BPNN and RBFNN will be utilized to compare with AFNN in the estimation of elbow joint angle. In [14] and [15], the optimal parameters of hidden layer neurons and input order have been discussed in the human lower limb joint angle estimation. Therefore, the optimal parameters of hidden layer neurons 20 As can be seen from the Fig.…”

“…[2], [7], [8]. In recent years, with the widespread application of neural networks in the fields of science and engineering [9]- [11], domestic and foreign scholars have proposed a variety of schemes, such as support vector machine (SVM) [12], artificial neural networks [13], BPNN [14] and RBFNN [15] etc, to estimate the joint angle from the surface biological signals, so as to realize the recognition of the human body intention. Specifically, In [16], Y.Masahiro et al proposed a continuous hand pose estimation method, which is based on the SVM and relationship between EMG signal and joint angle, and the experiments show a high accuracy rate for motion classification and joint angle estimation.…”

“…Simultaneously, the parameters discussion ensures that the RMS error of intention recognition is minimized under the optimal parameters. In addition, in [15], the RBFNN was utilized to estimate the joint angles of hip, knee, and ankle from the sEMG signals of rectus femoris, lateral femoral muscle and extensor halluces. The RMS error of intention recognition reached a controllable range.…”

A novel method based on long short term memory network and discrete-time zeroing neural algorithm for upper-limb continuous estimation using sEMG signals

“…The collection process was simple and non-invasive, making it a popular research field in human-machine interaction technology. At present, a fair amount of representative research on motion intention recognition based on sEMG has been reported, which can be broadly divided into two categories: classification of the motion pattern [9,10,11,12,13,14] and estimation of the continuous motion of the joint [15,16,17,18,19]. Many classification methods applied continuous sEMG signals to estimate joint angles with predefined sets, such as support vector machine [9], artificial neural network [10], gaussian mixture model [11] and other classifiers [12,13,14].…”

LS-SVM Combined with ZNN for Predicting the Continuous Motion Joint Angle of Lower Limb

“…This feature lays the foundation for the prediction of movement intention. Since the sEMG signal contains the command information of the neuromuscular system in real time, the technology of using sEMG to realize human-machine interaction has become the object of many researchers [8][9][10].…”