Slip suppression of prosthetic hand
Herein we present a multi-threshold-based constant micro-increment control strategy to detect and suppress the slip for the prosthetic hand, and to minimize the loading force increment after the stabilization. The proposed strategy primarily encompasses slipping process model, multi-threshold detection method, constant micro-increment controller and a preset filter. First and foremost, a slipping process model is proposed that involves the nonlinear and noise characteristics of the system. Then, to improve the detection accuracy, the multi-threshold is determined by the method of partitioned statistics to reduce the influence of the loading force and nonlinearity of the system on the detection part. At the same time, to achieve the goal of minimizing the loading force increment after stabilization, a constant micro-increment controller is employed. Finally, considering that the noise of the sensor plays a decisive role in the detection accuracy, a moving average filter is used as the preset filter and the effect is compared with low-pass filters. In the experiment, the parameters selection merely depends on the system characteristics and is basically irrelevant to the object being grasped. The experimental results show that the detection accuracy of the proposed algorithm is better than a single threshold detection algorithm. The time of slip prevention is less than 90ms for all the experiments, and the loading force increment is also comparatively small.
There are two sets of experiments, a total of four videos, which need to be decompressed and watched.