Wyniki wyszukiwania - BazTech

1

Evaluation of simple microphone-based mechanomyography (MMG) probe sets for hand stiffness classification

Zubrycki Igor, Granosik Grzegorz

Journal of Automation Mobile Robotics and Intelligent Systems

|

2019

|

Vol. 13, No. 2

28--39

EN

We describe simple to build mechanomyography sensors, with one or two channels, based on electret microphones. We evaluate their application as a source of information about the operator’s hand stiffness, which can be used for changing a robot’s gripper stiffness during teleoperation. We explain a data acquisition procedure for further employment of a machine-learning. Finally, we present the results of three experiments and various machine learning algorithms. support vector classification, random forests, and neural-network architectures (fullyconnected articial neural networks, recurrent, convolutional) were compared in two experiments. In first and second, two probes were used with a single participant, with probes displaced during learning and testing to evaluate the influence of probe placement on classifcation. In the third experiment, a dataset was collected using two probes and seven participants. As a result of the singleprobe tests, we achieved a (binary) classification accuracy of 94 % during the multi-probe tests, large crossparticipant differences in classifcation accuracy were noted, even when normalizing per-participant.

2

Klasyfikacja sztywności dłoni podczas teleoperacji za pomocą sondy mechanomiograficznej

Zubrycki I., Granosik G.

Prace Naukowe Politechniki Warszawskiej. Elektronika

|

2018

|

z. 196

523--534

PL

W artykule przedstawiamy ekonomiczny zestaw do mierzenia sygnałów mechanomiograficznych, wykorzystujący mikrofon elektretowy umieszczony w zaprojektowanej do tego celu obudowie. Zestaw został skonstruowany w celu określenia sztywności dłoni operatora, który za pomocą rękawicy sensorycznej steruje manipulatorem. Urządzenie stanowi uzupełnienie systemu do teleoperacji chwytakami zręcznymi [25, 24], dostarczając dodatkowych informacji o stanie sztywności ręki operatora, co można wykorzystać do ustalania impedancji chwytaka wielopalczastego. Przedstawiamy budowę urządzenia, sposób przetwarzania sygnału oraz porównujemy algorytmy uczenia maszynowego pozwalające na wykorzystanie urządzenia do rozpoznawania sztywności dłoni. Efektem pracy systemu jest binarne rozpoznawanie sztywności (sztywny, rozluźniony) z jakością klasyfikacji 94% przy dowolnym ruchu dłoni. Zaprojektowane urządzenie oraz algorytmy udostępnione są na otwartej licencji i dostępne są w repozytorium projektu.

EN

The article presents a low cost set for measuring mechanomyographic signals to determine the stiffness of the operator's hand during remote control of manipulator with a dextrous gripper. The device complements the system for teleoperation initially containing sensor glove and LeapMotion vision, and now providing additional information about the operator's hand condition, which can be used to determine the impedance of a multi-fingered gripper. We present the construction of the device, signal acquisition, and processing, the machine learning algorithm that provides the main functionality of the device: recognizing hand stiffness without the need for calibration with any external measurement tools. The output of the system is the binary recognition of the operator's hand stiffness (rigid, non-rigid) with the maximum quality of classification of around 94% regardless of sensor placement or movement of the hand.

3

Electromyography and mechanomyography signal recognition: Experimental analysis using multi-way array decomposition methods

Wołczowski A., Zdunek R.

Biocybernetics and Biomedical Engineering

|

2017

|

Vol. 37, no. 1

103--113

EN

In this study, we considered the problem of controlling a prosthetic hand with noisy electromyography (EMG) and mechanomyography (MMG) signals. Several dimensionality reduction methods were analyzed to assess their efficiency at classifying these signals, which were registered during the performance of grasping movements with various objects. Using the cross-validation technique, we compared various dimensionality reduction methods, such as principal components analysis, nonnegative matrix factorization, and some tensor decomposition models. The experimental results demonstrated that the high-est classification accuracy (exceeding 95% for all subjects when classifying 11 grasping movements) and lowest computational complexity were obtained when higher-order singular value decomposition was applied to a multi-way array of multi-channel spectrograms, where the temporal EMG/MMG signals from all channels were concatenated.

4

Verification of the functionality of device for monitoring human tremor

Szumilas M., Lewenstein K., Ślubowska E.

Biocybernetics and Biomedical Engineering

|

2015

|

Vol. 35, no. 4

240--246

EN

Tremor accompanying the Parkinson's disease is perceived as one of its most disturbing symptoms. Among available treatments there is a deep brain stimulation, which effectively reduces unwanted oscillations of patient's muscles. Nevertheless, setting parameters of the stimulation is a highly empirical process and the final outcome depends primarily on the experience of involved medical personnel. We present a device which is meant to provide a clinician with feedback based on the measurable parameters of tremor, monitored in many points of the body simultaneously. Functionality of the device was verified at a basic level. During the verification, the vibrations were recorded: (1) in a relaxed arm, (2) during voluntary contraction of muscles and (3) after being damped by tissues (in this case the vibrations were introduced from an external generator). Moreover, a method of selecting optimal place for mounting vibration probes is presented.

5

Characterization of the mechanomyographic signal of three different muscles and at different levels of isometric contractions

Jotta B., Garcia M. A. C., Pino A. V., Souza M. N. de

Acta of Bioengineering and Biomechanics

|

2015

|

Vol. 17, no. 4

73--84

EN

Purpose: Lateral (X) and longitudinal (Y) mechanical oscillations of muscle fibers that take place during muscular contraction seem to contain information additionally to the myoelectric activity, which can contribute to the interpretation of some muscle gradation force mechanisms. However, no previous study was found that had investigated the relationship between the muscle force and features associated to the mechanomyographic (MMG) signal obtained by means of a biaxial accelerometer in three different muscles. Therefore, the aim of this study was to evaluate the relationship between the force output at different load levels (20% to 100%) of the maximum voluntary isometric contraction (%MVIC) and the two signals supplied by a biaxial accelerometer and, in addition, the so-called resultant (R) acceleration signal derived from the two signals mentioned previously. Twenty seven male volunteers participated in this study. Methods: The force output related to the right biceps brachii, soleus and gastrocnemius medialis muscles was studied by means of linear regression models fit to log-transformed of the root mean square (RMS) values of the MMG signals in X, Y, and R axes versus each %MVIC. The phase angle of R acceleration (PhaseR) and anthropometric data were also considered. Results: The angular coefficient a and the antilog of y-intercept b from the log-transformed of MMG data values versus force output were able to distinguish partially motor unit strategies during isometric contractions in the three muscles studied. Conclusion: The findings suggest that biaxial accelerometer seems to be an interesting approach in the assessment of muscle contraction properties.