Comparison of Root Mean Square Index and Hilbert-Huang Transform for Detection of Muscle Activation in a Person with Elbow Disarticulation

An active prosthesis is a device developed to substitute an absent limb of the human body supplying its functionalities without neglecting the patient's body image. The development of these devices includes the signal acquisition process, signal conditioning, onset/offset detection, classification algorithms that determine which movement is being realized, and actuators that executed the motion. One of the important aspects, and with few studies are the techniques of onset/offset detection which allow performing the movement with the minimum delay and stop executing the movement to then proceed to the classification stage. Normally, this stage is not considered due to the design of the prosthesis a part of the signal is taken by suppressing the start and the end of the contraction considering only the isotonic component. In this study, the comparison of two techniques, the root means square index and the Hilbert-Huang Transform concerning the visual inspection technique for the onset/offset detection is realized. Statistical analysis is performed by considering the median, maximum value, minimum value, mean, standard deviation, and dispersion of the data to determine which technique is more adequate. Additionally, a protocol is proposed to acquire signals for the performance of 6 movements: elbow flexion, elbow extension, pronation and supination of the forearm, and opening and closing of the hand. The study was performed on a person with elbow disarticulation and a healthy person. The results indicate that the most appropriate technique for onset/offset detection in a person with elbow disarticulation is the Hilbert-Huang transform indicating greater accuracy.