| As a kind of human organs, hand is one of the most commonly used in our daily life. And it is the highest embodiment of the human body movement dexterity. However the stroke can raise central nerve damage and lead to obstacle of hand movement function, which may affect their daily life. Neuroscience has proved that rehabilitation training can help patients to restore movement function. With the development of the rehabilitation robot technology at home and abroad, the robot assisted therapy has become a main means of present rehabilitation treatment.This paper first analyzed the status of hand structure and rehabilitation robot from the view of physiology. Then we summarized the advantages and disadvantages of several kinds of rehabilitation strategy used in robot therapy. After that we expounded the research plan and design in our project. Based on the analysis of skeletal muscle physical model of the finger, a kind of hand parametric structure model is established. Combining with the laboratory research and development of hand rehabilitation robot hardware platform, we design a method that can obtain finger muscle force and parameters.In order to implement the active strategy of hand rehabilitation robot effectively, this paper constructed a variable impedance control system based on fuzzy neural network, which can realize the robot rehabilitation training based on the active strategy. The impedance control system was designed to collect the information such as muscle force and motion angle. And the impedance parameters were obtained from the real-time system of fuzzy neural network based on an off-line learning system. Finally the robot trajectory tracking was completed through the control system.In addition, a human-computer interaction platform was designed based virtual simulation, which can be used for manipulation of the robot system with the patient’s rehabilitation training. Interactive software was run on the Android system. According to the rules of life cycle of the Android activity and data transfer, a data communication system was established. On the basis of kinematic analysis, the virtual simulation system is constructed to a real-time feedback on rehabilitation training and the goal of the movement. Moreover, some rehabilitation training scenarios was designed to induce patients to participate in active rehabilitation training.The whole hand rehabilitation robot experiment platform was set up finally. The stability of the whole robot system was optimized and the impedance of the fuzzy neural network control system based on the parameterized hand modeling was verified. Then data communication system and virtual simulation environment was tried to run on the human-computer interaction platform based on Android. Hand rehabilitation robot system was combined in the final experiment. Experiments proved that the system can be stable operation, and the results also proved that this system can realize the hand active rehabilitation training. |
