A Novel TCN-LSTM Hybrid Model for sEMG-Based Continuous Estimation of Wrist Joint Angles

Abstract

Surface electromyography (sEMG) offers a novel method in human-machine interactions (HMIs) since it is a distinct physiological electrical signal that conceals human movement intention and muscle information. Unfortunately, the nonlinear and non-smooth features of sEMG signals often make joint angle estimation difficult. This paper proposes a joint angle prediction model for the continuous estimation of wrist motion angle changes based on sEMG signals. The proposed model combines a temporal convolutional network (TCN) with a long short-term memory (LSTM) network, where the TCN can sense local information and mine the deeper information of the sEMG signals, while LSTM, with its excellent temporal memory capability, can make up for the lack of the ability of the TCN to capture the long-term dependence of the sEMG signals, resulting in a better prediction. We validated the proposed method in the publicly available Ninapro DB1 dataset by selecting the first eight subjects and picking three types of wrist-dependent movements: wrist flexion (WF), wrist ulnar deviation (WUD), and wrist extension and closed hand (WECH). Finally, the proposed TCN-LSTM model was compared with the TCN and LSTM models. The proposed TCN-LSTM outperformed the TCN and LSTM models in terms of the root mean square error (RMSE) and average coefficient of determination (R²). The TCN-LSTM model achieved an average RMSE of 0.064, representing a 41% reduction compared to the TCN model and a 52% reduction compared to the LSTM model. The TCN-LSTM also achieved an average R² of 0.93, indicating an 11% improvement over the TCN model and an 18% improvement over the LSTM model.

Keywords: human–machine interaction (HMI); long short-term memory neural network (LSTM); surface electromyography (sEMG); temporal convolution network (TCN); wrist kinematics estimation.

Figure 1

The equipment for recording experimental data.

Figure 2

The target angle we selected for experiments.

Figure 3

The three types of movement we selected: (a) wrist flexion (WF); (b) wrist ulnar deviation (WUD); (c) wrist extension and closed hand (WECH).

Figure 4

Convolution architecture of a TCN.

Figure 5

Residual connection block of the TCN.

Figure 6

The architecture of the LSTM network unit.

Figure 7

The architecture of the hybrid TCN-LSTM model.

Figure 8

The relationship between RMSE and channel number and kernel size.

Figure 9

Estimation results of WECH movement in eight subjects using TCN-LSTM model.

Figure 10

Estimation results of WECH movement in eight subjects using TCN model.

Figure 11

Estimation results of WECH movement in eight subjects using LSTM model.

Figure 12

The histogram of the average performance for the selected movement types in the 8 subjects. (a) Average RMSE results for the three movements using the three models. (b) Average R² results for the three movements using the three models. ‘*’ indicates statistical significance (p < 0.05).

Figure 13

Comparison of training time of the three models.

Figure 14

Histogram of the average performance of 5 models for different movement types. (a) Average RMSE results of five models for WF, WUD, and WECH. (b) Average R² results of five models for WF, WUD, and WECH.