As robots become increasingly integrated into human society, the importance of human–machine interfaces continues to grow. This study proposes a faster and more accurate control system for myoelectric prostheses by considering the Electromechanical Delay (EMD), a key characteristic of Electromyography (EMG) signals. Previous studies have focused on systems designed for wrist movements without attempting implementation. To overcome this, we expanded the system’s capability to handle more complex movements, such as those of fingers, by replacing the existing four-channel wired EMG sensor with an eight-channel wireless EMG sensor. This replacement improved the number of channels and user convenience. Additionally, we analyzed the communication delay introduced by this change and validated the feasibility of utilizing EMD. Furthermore, to address the limitations of the SISO-NARX model, we proposed a MISO-NARX model. To resolve issues related to model complexity and reduced accuracy due to the increased number of EMG channels, we introduced ridge regression, improving the system identification accuracy. Finally, we applied the ZPETC+PID controller to an actual servo motor and verified its performance. The results showed that the system reached the target value approximately 0.240 s faster than the response time of 0.428 s without the controller. This study significantly enhances the responsiveness and accuracy of myoelectric prostheses and is expected to contribute to the development of practical devices in the future.

Sensor re-donning can alter surface electromyography (sEMG) amplitudes and inter-channel relationships, degrading sEMG-based motion classifiers. This study proposes personalized recalibration for four-finger motion classifiers and evaluates whether reusing prior Training data can reduce newly acquired Calibration data. Limited labeled Calibration data transform Training signals toward the current condition via classwise amplitude correction and classwise regularized covariance alignment before support vector machine retraining. Three able-bodied male participants (P1–P3) were evaluated using Stable Motion window-level accuracy. For P1, accuracy pooled over ten Test datasets from five re-donning conditions increased from 88.90% to 98.88%. For P2 and P3, exploratory post hoc analyses of four failure-case directions per participant showed pooled accuracy increasing from 24.46% to 98.97% and from 73.74% to 96.72%, respectively. Across exhaustive retrospective analyses of all shared two-of-five repetition subsets, Combined accuracy was 98.29%, 98.68%, and 95.58% for P1–P3, respectively, only 0.18–0.58 percentage points below the five-repetition results; prospectively shortened Calibration recordings were not evaluated. These findings support the feasibility of participant-specific recalibration for Stable Motion four-finger classification after sensor re-donning and warrant further evaluation in broader user populations and continuous-operation settings.

1. A global burden analysis estimated that 57.7 million people worldwide were living with limb amputation due to traumatic causes in 2017 [1]. Another epidemiological study reported that, from 1990...