**2. Procedure of EMG detection**

a replacement for the lost body part [5, 8]. The meal assistive robot [9], assistive wheel chair [10] and assistive humanoid robot [10] are some examples of the other types of assistive robots. Both orthoses and prostheses directly interact physically and cognitively with the wearer. Therefore, they are expected to provide physiological and mental comfort to the user, without letting the user feel a difference while functioning/assisting correctly to perform the required motion. However, the controlling of the robot according to human motion intention is not an easy task [11, 12]. Therefore, in the integration of the human with a robot, the selection of a proper control input signal to reflect the correct motion intention would be very important. So far research is being carried out considering different biological signals such as Electromyog‐ raphy (EMG), Mechnanomyogram (MMG), Electroencephalography (EEG) Electrooculogra‐ phy (EOG) and Electrocorticogram (EcoG) [13, 14, 15] as the main input signal to the robot controller. Among them, the EMG signal, which is the measurement of the electrical activity of muscles at rest and during contraction, has obtained promising results in the case of controlling robotic prostheses and orthoses by correctly interpreting the human motion intention [16, 17]. Basically, EMG based control is a sophisticated technique concerned with detection, processing, classification and application of EMG to control assistive robots. EMG can be applied to control assistive robots in various manners by considering data acquisition, feature extraction and classification [18]. Some of the available exoskeleton robots controlled based on EMG are the orthotic exoskeleton hand [19], exoskeleton robot for tremor suppression [20], SUEFUL-7 [6] etc. In addition, DEKA Arm [21], Saga Prosthetic Arm [22], and Manus Hand [23] are some of the available prosthetic devices with EMG based controlling. Many researchers from different institutions have been working on EMG based control methods over several decades and their contribution has greatly influenced a new era of EMG based control

This chapter presents a comprehensive review of EMG based control methods of recent upperlimb orthoses and prostheses. Initially, it explains the detection and processing of EMG and available EMG extraction systems. Next, ways of categorization of EMG based control methods are explained. Here EMG based control methods are categorized in detail, based on the structure of the controller algorithm, as pattern recognition based controls and non-pattern recognition based controls. Most of available control methodologies used with assistive robots are pattern recognition based controls [18, 19, 24, 26, 27]. The control methods belonging to non-pattern recognition based controls [28-30] are rare. Further, comparison of EMG based control methods of upper-limb exoskeleton robots and upper-limb prostheses are presented considering the features of the control method. In addition, the conclusion and future direc‐

The next section focuses on the procedure of EMG detection and processing. Section 3 describes the categorization of EMG based control methods. The review of EMG based control methods of upper-limb exoskeleton robots and upper-limb prostheses together with a comparison of control methods are presented in section 4. The final section briefly outlines the conclusion

tions of EMG based control methods of assistive robots are described.

of assistive robots [5, 24, 25].

238 Electrodiagnosis in New Frontiers of Clinical Research

and future directions.

Detection of EMG signals can be done mainly in two ways, namely non-invasive and invasive [5]. Surface EMG (sEMG) electrodes are used for the former, while intramuscular EMG (iEMG) electrodes are used for the latter. Placement of surface EMG electrodes is comparatively easier than intramuscular EMG electrodes. However noise and other disturbances are inherent in surface EMG detection [5]. Intramuscular EMG electrodes are placed close to the Motor Unit Action Potentials (MUAP), and as a result the influence of other disturbances is not dominant. It provides a better accuracy and repeatability of the EMG signal [25]. As shown in Fig. 1, the EMG extraction process includes several steps. The initial step is the selection of the most significant muscle of the human body relevant to the required motion. After the muscle is selected, the next important step is the placement of electrodes. In the case of sEMG, the electrodes should be placed in the belly area of the muscle for maximum signal extraction.

**Figure 1.** EMG signal extraction process

The electrode should be placed onto the relevant muscle after cleaning the skin surface. There are a few types of surface electrodes, some of which need a gel [31] to be applied between the skin and the electrode and some [32] which instead use an adhesive tape to ensure proper contact between the muscle and the electrode. Signals from several electrodes are then fed into the input box and subsequently passed to the amplifier. The output of the amplifier can be fed into a computer via data cards or any other data communication interface and can be recorded or manipulated in the required way. In most EMG based control methods a raw EMG signal is processed to extract the features of the signal. Several feature extraction methods are available for this purpose [5]: mean absolute value, mean absolute value slope, waveform length, zero crossings, root mean square value, *etc*. The features of the raw EMG signal have to be extracted in real time to use EMG as input signals to the controller of the assistive robots. Most robots use Root Mean Square (RMS) as the feature extraction method of raw EMG mainly due to ease of analyzing real time information of EMG signal.

There are a number of commercially developed EMG acquisition systems available [31-34] (see Fig. 2). They could be used for both medical and research purposes. Some of the leading EMG acquisition system manufacturers are Nihon Kohden Co. [31], Delsys [32], BioSemi [33], and Cambridge Electronic Design [34]. The Delsys EMG system shown in Fig. 2(a) is widely used in research, whilst the Nihon Kohden EMG system shown in Fig. 2(b) is widely used for medical applications. Nevertheless, there are other models in Nihon Kohden which also support for research [31]. The BioSemi EMG system shown in Fig. 2(c) could also be used in research applications. The next section will explain the classification of EMG based control methods

(a)Delsys [32] (b)Nihon Kohden[31] (c)Bio Semi [33]

Data Segmentation

Data

**Figure 5.** EMG processing sequence in non-pattern recognition based control

**3. Categorization of EMG based control methods**

**Figure 4.** EMG processing sequence in pattern recognition based control

acuqisition

acuqisition

(see Fig.5).

Feature

Recent Trends in EMG-Based Control Methods for Assistive Robots

http://dx.doi.org/10.5772/56174

241

Extraction Classification EMG

Segmentation Non pattern recognition method EMG

Control methods of assistive robots based on EMG can be categorized mainly according to the input information to the controller, architecture of the control algorithm, output of the controller and other ways. Figure 3 shows the ways of categorization of EMG based control methods. In this chapter, the EMG based control methods of assistive robots are categorized based on the architecture of the control method. Considering the EMG signal processing method in the architecture of the controller the EMG based control methods can be categorized mainly as pattern recognition based and non-pattern recognition based [5, 25]. Control methods of many assistive robots are designed with pattern recognition based control methods and it provides an accurate control action compared to non-pattern recognition based EMG based control [5]. However, several intermediate steps (see Fig.4) are applied with a pattern recognition method such as data segmentation, feature extraction, and classification [5, 25]. The accuracy of pattern recognition based control is greatly improved by methods of feature extraction and classification [5, 25]. Robots such as, SUEFUL-7 [6], NEUROExos [27], W-EXOS [11], DEKA Arm [21], Saga Prosthetic Arm [22], Manus Hand [23] are on pattern recognition based control. The non-pattern recognition based control method involves only a few steps

**Figure 2.** Available EMG acquiring systems

**Figure 3.** Control method classification for assistive robots

**Figure 4.** EMG processing sequence in pattern recognition based control

(a)Delsys [32] (b)Nihon Kohden[31] (c)Bio Semi [33]

Control Model Input Output

> Dynamic model Muscle model On off control Proportional control Proportional+Derivative

Fuzzy control Neuro-fuzzy control Impedance control Admittance control

Proportional+Derivative+Integral

Current

Voltage

**Figure 2.** Available EMG acquiring systems

240 Electrodiagnosis in New Frontiers of Clinical Research

Biological signal (EMG/EEG/EOG/MMG)

Non-biological signal (Force/Torque signal)

**Figure 3.** Control method classification for assistive robots

**Figure 5.** EMG processing sequence in non-pattern recognition based control
