EMG control of hand motion

Detecting the EMG signals, the electrodes usually have to be pasted on the agonist muscle. The selection of the pasting position requires bio-mechanical knowledge, which will prevent the EMG signals from spreading as a new human interface against the mechanical devices. To solve this problem, we examine a method to place the multiple electrodes around the limbs, more preciously, a fore arm, to estimate the hand motions. Applying the signal processing such as primary component analysis, independent component analysis, neural networks, or support vector machines, we attempt to estimate the agonist signals, or hand motion direcctlly.