Introduction:

the paper presents the prediction of three basic hand movement types by means of a smart algorithm to draw characteristics indispensable for identification of movement patterns based on the analysis of surface electromyographic signals obtained with the Myo device.

Objective:

recognize and predict basic movement patterns of the arm joint using surface electromyography with a view to applying them over a prosthesis prototype.

Methods:

data were taken from 13 students aged 22 and 23 years from the Salesian Polytechnic University, each of whom performed three types of grasp: cylindrical, pincer and palmar pincer. A 10 Hz frequency was used and 5 samples were taken of each grasp type during 60 seconds. Statistical analysis was performed with the tool ANOVA, establishing a significance value > 0.65.

Results:

in certain volunteers a greater reaction was observed in electrode 1, due to their larger forearms. Response time for identification varies with the number of variables to be compared. When only one movement is analyzed, response time is 2.6 seconds, but when the three movements are examined it rises to 7.8 seconds by the number of electrodes intended to be studied.

Conclusions:

the response of the system proposed starts to slow down as more movements are analyzed simultaneously, which makes it less effective. The performance and response time of our system is higher than in state-of-the-art systems, since fewer signal characterization methods are used. On the other hand, a limitation of the project is the sampling frequency of the Myo device (200 Hz).

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )