Disturbance-Observer-Based Fuzzy Control for a Robot Manipulator Using an EMG-Driven Neuromusculoskeletal Model

Figure 1

Schematic structure of the controller framework based on an EMG-driven NMS model. The NMS model consists of four components: (A1) the model’s musculotendon kinematics were used to compute musculotendon lengths and moment arms; (A2) muscle activation dynamics were employed to calculate the level of muscle activation involved in the processed EMG signals; (A3) muscle contraction dynamics according to a Hill-type muscle model were applied to predict musculotendon force, using calculated musculotendon length and muscle activation as inputs; (A4) joint dynamics was used to compute joint torques, using the calculated musculotendon forces and moment arms as inputs. The EMG signals and joint angles from the user were used to estimate the motion intention (desired torque) through the EMG-driven NMS model. Then, the desired torque was transmitted to the desired velocity through an admittance filter. A disturbance-observer-based adaptive fuzzy controller was developed for the robot manipulator system with model uncertainties.