For example, when performing leg flexion at standing position (STD), the gluteus maximus, quadriceps, triceps surae and erector spinae are recruited to keep balance for straight standing, while femoral biceps (FB) and semitendinosus (SEM) are recruited as agonistic muscles with rectus femoris (RF) and vastus medialis (VM) as antagonist muscles for knee flexion. Each voluntary motion involves the activation of a large amount of muscles, and the functions of muscles vary from motion to motion. The sEMG sensors are used to record and analyze muscle activation during gestures and motions. Compared with these methods, surface electromyography (sEMG) is a kind of portable, non-invasive and low-cost wearable device, which has been widely used for assessing human activities, especially in the field of neurological or musculoskeletal disorder diagnosis, exoskeleton control and rehabilitation guidance. Traditional technologies for motion identification and disease assessment require expensive laboratorial instruments, such as gait channel, electro-optical motion capture and three-dimensional force platform. By analyzing the sensor data during motions, we can establish an objective method to identify a patient’s movement intention or disease state, and this has been widely applied in fall detection, artificial limbs and exoskeletons, as well as disease monitoring and assessment. Recently, sensor-based intelligent motion recognition and disease diagnosis has been greatly developed due to the rapid development of sensing, machine learning and computing science. Lower limb motions are essential in many human activities, especially the ability to independently perform the activity of daily living (ADL), such as standing, sitting and walking, and very important for avoiding injuries caused by falls, improving personal happiness and reducing the social costs of nursing. Results show that the NMF-based sEMG classifier can achieve good performance in lower limb motion recognition, and is also an attractive solution for clinical application of knee pathology diagnosis. The study was conducted on an open dataset of 11 healthy subjects and 11 patients. Finally, the random forest classifier is adopted to identify motions or diagnose knee pathology. Then, a two-stage feature selection method is executed to reduce the dimension of feature sets extracted from aforementioned matrices. First, raw sEMG segments are preprocessed and then decomposed to muscle synergy matrices by NMF. Based on that, we have designed a classification method for motion recognition and knee pathology diagnosis. The similarities of muscle synergy of subjects with and without knee pathology performing three different lower limb motions are calculated.
To solve this, muscle synergy analysis based on non-negative matrix factorization (NMF) of sEMG is carried out in this manuscript. However, due to the complex nature of neural control in lower limb motions and the divergences in subjects’ health and habits, it is difficult to directly use the raw sEMG signals to establish a robust sEMG analysis system. Surface electromyography (sEMG) has great potential in investigating the neuromuscular mechanism for knee pathology.
0 kommentar(er)
