The collection of myoelectric information, including the detection and guidance of myoelectric information, display and recording, and special acquisition techniques, such as electrode fixation and positioning technology. EMG signals are usually affected by the following factors: 　　 (1) the type, function and condition of the muscle ( including fatigue ); 　　 (2) the characteristics of tissues, bones and skin located between the muscles and the electrodes ; 　　 (3) Electrode material, surface structure, geometry and spacing: 　　 (4) The position of the electrode relative to the skin. Our safes are divided into mechanical lock safes, electronic password safes, fingerprint safes and other different panels. In terms of material, we are all made of alloy steel. The thickness of the door plate is 1.2-10mm, and the thickness of the box plate is 1.2-6mm. We pay special attention to the anti-theft function. Each safe is equipped with three-sided solid lock bolts to increase the stress area on the side and effectively realize anti-pry and, house hold safe with a height of 35cm or above and each safe box is equipped with multiple customized hooks inside, which is beautiful and practical, and improves the utilization rate of space. In addition, each safe is equipped with shock-absorbing and moisture-proof interior decoration to better protect valuables and personal belongings from scratching. Safes with a height of 45cm and above are equipped with removable partitions, so you can freely allocate your storage space.Now we are looking for more partners worldwide. If you are looking for safes of any size, you can contact us online and sincerely hope to get your attention and cooperation opportunities. Tiger Safe Box,LCD Electronic Safe,Gold Color Safe Box,Digital Steel Security Safes,Safe Deposit Box Hebei Hupai Aodi Cabinet Industry Co.,Ltd. , https://www.hupaiaodi.com
Introduction to factors affecting the collection of myoelectric signals and methods for processing myoelectric information
The detection of human bioelectricity must rely on the guidance of a sensor-electrode. Since the myoelectricity is weak in human bioelectricity, the electrode selection for guiding myoelectricity is very large. Commonly used are surface electrodes and needle electrodes (see Table 1 for comparison). The size, shape, material, and process of the electrode have an effect on the performance of the detected myoelectricity. The electrodes are usually made of silver sheets, and materials such as gold and aluminum are also used. The shape of the surface electrode is generally circular, and the diameter is about 7 to 10 coffees, and the large ones are up to 15 mm.
Regardless of the electrode used, the detected signal is the vector sum of the myoelectric potentials emitted by many motor units, and the EMG signals of different actions must be separated by appropriate analysis methods. This is the core of myoelectric prosthesis, which directly affects the accuracy and flexibility of prosthetic control.
The current EMG information processing technologies mainly include the following:
(1) Conventional time domain amplitude electromyography treatment method
This method uses the mean value of the time-domain peak-to-peak value of the myoelectric signal, or the integrated characteristics of the muscle signal collected by the whole set of electrodes, and controls the corresponding artificial hand motion after being recognized by the control circuit. It is characterized by a pair of electrodes or a domain voltage that can only control one degree of freedom. Most commercial prosthetics currently use this control strategy. If you want to control multiple degrees of freedom, you must have several pairs of electrodes to collect the signal. It is difficult to find several pairs of muscles that can meet the multi-degree of freedom control requirements in the amputee, and the multi-electrode will greatly reduce the control. Reliability and stability.
(2) Time series analysis method
This method considers the multi-channel of surface EMG signal as an autoregressive process of vector values, and uses time series method to extract the information needed for limb function classification, and obtain the prediction of the value and direction of limb movement. The EMG signal can be equivalent to the output of a linear system excited by a zero-mean white noise process. A typical 4-6 order AR model can accurately classify the EMG signals of the upper limbs. In recent years, the study of EMG mainly uses two-channel signals, and the AR model is established by Marpie algorithm to obtain AR parameters, and then the Bayesian criterion is used to discriminate the actions. Others use the Hopfield network to extract AR coefficients.
However, due to the presence of non-static factors in the EM6 signal, it is difficult for the AR model to accurately extract feature parameters. For the same reason, non-static processing models such as adaptive models are not very good, mainly due to the difference in estimates generated and parameter oscillations. Kivyu et al. found that the AR parameter has time-varying characteristics when used as a linear force force on the muscle, which results in the motion potential of the motion unit becoming unclear under broadband noise and low amplitude, especially at low levels. The forced contraction is more obvious.
(3) Neural network processing method:
Neural networks have significant persistence when dealing with complex, fuzzy, noisy, and unsteady signals, and are adaptive to gradual changes in EMG signals. It can also replace most of the work in the rehabilitation training, which effectively simplifies the control operation and improves the control efficiency. Therefore, neural networks are very suitable for EMG signal analysis.
In recent years, the number of documents using neural networks to process myoelectric information is not uncommon. For example, in the literature o" as a classifier for the spectrum of myoelectric signals, in the literature "" is used to obtain the nonlinear mapping relationship between the first chapter of the electromyography information and the prosthetic action. The document "" constructs an ANN. A multi-layered visual sensor and an ANN sifter using a new non-monitoring culture strategy to obtain the shape of each motion action potential waveform. The literature uses a multilayer perceptron and a hidden Markov model to combine the EMG signals. Classification can not only solve the defect of poor recognition, but also consider the dynamic performance of the FAJG signal.
In addition, many scholars have also proposed other methods of processing EMG information. For example, the evaluation of the rate of change of myoelectric signals, using fast Fourier analysis, heuristic rule method, spatial factor method, fuzzy classification method, statistical pattern recognition method, wavelet transform method, histogram recognition method, etc., have achieved good results. Another common method is to combine the neural network and the AR model to process the EMG. For example, the AR model is used to detect the characteristic parameters of the EMG, and the neural network is used to complete the pattern classification of the specific motion or potential motion of the arm; the literature distortion 1 uses the four parameters of the AR model and the power of the EMG signal as the input of the multilayer perceptron network. , can control six degrees of freedom. The success rate of the above methods has reached or exceeded 95%.
In summary, although the myoelectric prosthesis has certain defects, its unique advantages still make it a hot spot in the research of prosthetic limbs. The commercialized fake hands are only mature, and how to improve their control accuracy. The rate is the focus of the myoelectric control fake hand. This article will continue to explore the single-degree-of-freedom myoelectric artificial hand. The main work done in this paper is as follows:
1) The design of the EMG signal extraction device, that is, the design of the electrode fabrication and extraction circuit:
2) Preparation of the EMG signal analysis program;
3) The design of the artificial hand control circuit, that is, the design of the preamplifier circuit and the drive circuit.