Abstract:The traditional classification methods for limb motion recognition based on sEMG have been deeply researched and shown promising results. However, information loss during feature extraction reduces the recognition accuracy. To obtain higher accuracy, the deep learning method was introduced. In this paper, we propose a parallel multiple-scale convolution architecture. Compared with the state-of-art methods, the proposed architecture fully considers the characteristics of the sEMG signal. Larger sizes of kernel filter than commonly used in other CNN-based hand recognition methods are adopted. Meanwhile, the characteristics of the sEMG signal, that is, muscle independence, is considered when designing the architecture. All the classification methods were evaluated on the NinaPro database. The results show that the proposed architecture has the highest recognition accuracy. Furthermore, the results indicate that parallel multiple-scale convolution architecture with larger size of kernel filter and considering muscle independence can significantly increase the classification accuracy.

In this article, a combined control strategy incorporating off-line iterative learning control and modified internal model control is proposed for improving the time waveform replication performance of electro-hydraulic shaking table. To reduce the modeling error between the estimated inverse model and the actual system, a modified internal model control strategy is first utilized to cope with the modeling error by back absorbing the nominal model and the inverse controller into a direct through block. Due to the non-minimum phase property of the nominal model estimated by the recursive extended least square algorithm, the zero magnitude error tracking controller is exploited to obtain a stable inverse controller. Then, an off-line iterative learning control strategy involving a real-time feedback controller is conducted on the compensated system to further enhance the replication performance. Therefore, the proposed algorithm combines the merits of modified internal model control and off-line iterative learning control and simplifies the conventional iterative control process by eliminating consecutive computation of Fourier and inverse Fourier transforms. The combined strategy is first programmed in MATLAB/Simulink and then compiled to a real-time personal computer with xPC target technology for implementation. Experiment results demonstrate that a better tracking accuracy and a faster convergence rate are achieved with the proposed algorithm than conventional pure iterative learning controller.

An electro-hydraulic shaking table is a useful experimental apparatus to real-time replicate the desired acceleration signal for evaluating the performance of the tested structural systems. The article proposes a combined control strategy to improve the tracking accuracy of the electro-hydraulic shaking table. First, the combined control strategy utilizes an adaptive inverse control as a feedforward controller for extending the acceleration frequency bandwidth of the electro-hydraulic shaking table when the estimated plant model may be a nonminimum phase system and its inverse model is an unstable system. The adaptive inverse control feedforward compensator guarantees the stability of the estimated inverse transfer function. Then, the combined control strategy employs an improved internal model control for obtaining high fidelity tracking accuracy after the modeling error between the estimated inverse transfer function using adaptive inverse control and the electro-hydraulic shaking table actual inverse system is improved by the improved internal model control. So, the proposed control strategy combines the merits of adaptive inverse control feedforward compensator and improved internal model control. The combined strategy is programmed in MATLAB/Simulink, and then is compiled to a real-time PC system with xPC target technology for implementation. The experimental results demonstrate that a better tracking performance with the proposed combined control strategy is achieved in an electro-hydraulic shaking table than with a conventional controller.

Rigid body orientation determined by IMU (Inertial Measurement Unit) is widely applied in robotics, navigation, rehabilitation, and human-computer interaction. In this paper, aiming at dynamically fusing quaternions computed from angular rate integration and FQA algorithm, a quaternion-based complementary filter algorithm is proposed to support a computationally efficient, wearable motion-tracking system. Firstly, a gradient descent method is used to determine a function from several sample points. Secondly, this function is used to dynamically estimate the fusion coefficient based on the deviation between measured magnetic field, gravity vectors and their references in Earth-fixed frame. Thirdly, a test machine is designed to evaluate the performance of designed filter. Experimental results validate the filter design and show its potential of real-time human motion tracking.

No abstract

No abstract

No abstract

No abstract