External control of movements and artificial neural networks

In this review paper we present the state of the art and suggest the solution that most likely could improve external control of movement by means of a Neural Prosthesis (NP). The model of the controller could be considered as a very simplified clone of the biological controller. External control that we advocate uses a hierarchical model that interfaces the biological systems at the level of command, and the level of execution. At the top level discrete sequential model of movement is implemented and relies on temporal synergies typical for biological control and sensors information. At the coordination level the decomposition to lower centers is implemented based on spatial synergies. At the actuator level customized biomechanical models need to be used in order to match the performances of the system. The top and coordination levels could be best understood as sets of rules having IF-THEN- ELSE form, therefore the Artificial Neural Networks (ANN) seem to be the most promising technique for their determination. In this review we describe the essence of the method, and list many of the applications of ANN for designing of the external control for NP. In addition, five presentations that follow discuss the mapping of the sensors information to the control of individual muscles, mapping of the sensors information for the synthesis of gait phases, use of ANN for determination of optimal electrode, application of wavelet to extract the important component of biological signals for control of movement) [41–45].