As an integral part of the body, the limb poses dexterous and fine motor grasping and sensing capabilities that enable humans to effectively communicate with their environment during activities of daily living (ADL). Hence, limb loss severely limits individuals' ability especially when they need to perform tasks requiring their limb functions during ADL, thus leading to decreased quality of life. To effectively restore limb functions in amputees, the advanced prostheses that are controlled by electromyography (EMG) signal have been widely investigated and used. Since EMG signals reflect neural activity, they would contain information on the muscle activation related to limb motions. Pattern recognition-based myoelectric control is an important branch of the EMG-based prosthetic control. And the EMG-based prosthetic control theoretically supports multiple degrees of freedom movements that allows amputees to intuitively manipulate the device. This chapter focuses on EMG-based prosthetic control strategy that involves utilizing intelligent computational technique to decode upper limb movement intentions from which control commands are derived. Additionally, different techniques/methods for improving the overall performance of EMG-based prostheses control strategy were introduced and discussed in this chapter.
Keywords: Limb motion; Myoelectric signal; Pattern recognition; Prosthesis; Rehabilitation robot; Upper limb amputee.