Combining Electromyography and Tactile Myography to Improve Hand and Wrist Activity Detection in Prostheses

Abstract

Despite recent advances in prosthetics and assistive robotics in general, robust simultaneous and proportional control of dexterous prosthetic devices remains an unsolved problem, mainly because of inadequate sensorization. In this paper, we study the application of regression to muscle activity, detected using a flexible tactile sensor recording muscle bulging in the forearm (tactile myography-TMG). The sensor is made of 320 highly sensitive cells organized in an array forming a bracelet. We propose the use of Gaussian process regression to improve the prediction of wrist, hand and single-finger activation, using TMG, surface electromyography (sEMG; the traditional approach in the field), and a combination of the two. We prove the effectiveness of the approach for different levels of activations in a real-time goal-reaching experiment using tactile data. Furthermore, we performed a batch comparison between the different forms of sensorization, using a Gaussian process with different kernel distances.

Special Issue on “Assistive Robotics”

“…The authors applied this solution to experimental data from healthy human subjects and amputee human subjects and found that the results from both group are comparable regarding classification accuracy. This work complements the work in [5]; both push the boundaries of experimental research in prosthetic device control and activation. Krings and Weinberger in [7] discuss the role and function of assistance in assistive robotics technologies for inpatient care solutions and debate the terms in the context of socio-technical systems.…”

“…They integrated human-generated feedback to obtain task engagement, which will be used for adjusting task parameters and difficulty level as demonstrated through their experimental results. The authors of [5] studied the problem of sensorization in dexterous control of prosthetic devices. Specifically, they focused on how regression models can be used to predict movement activation at various levels (wrist, hand, and single finger) from the tactile sensors.…”

Online Natural Myocontrol of Combined Hand and Wrist Actions Using Tactile Myography and the Biomechanics of Grasping

“…In this study, two feature selection methods were used in order to determine the prediction. The first one had already been used successfully in previous online studies with the first generation tactile bracelet (Jaquier et al, 2017;Nissler et al, 2017) and consists of the unprocessed data (except the Butterworth filter mentioned above) directly fed to a simple ridge regression (RR) algorithm. More precisely, the data consists of 288 filtered sensor data (9 boards of the 32 sensors each).…”

“…The term TMG is used for highdensity FMG: a technique in which many force/pressure sensors are put in contact with the subject's limbs. TMG has already been proved at least in Radmand et al (2016) and Jaquier et al (2017) and it has been shown to offer an unprecedented detail about the muscle patterns under examination. However, in these different works, combined motions were not tested.…”

Soft human–machine interfaces: design, sensing and stimulation

“…Fig.12(a) depicted the sensors attached to the wrist and elbow to record human deformation signals during the sporting process. Muscle activities were detected by recording muscle bulging in the forearm using a flexible tactile sensor, through which Gaussian process regression was carried out to predict wrist, hand and single-finger activation for controlling prosthesis and assistive robots [142]. Fig.12(b) depicted that thin carbon-black-doped PDMS was designed as the strain gauges featured with its high resistivity and strong dependence on strains, which was used for human gesture detection [34].…”