Abstract:
Robot-assisted rehabilitation solutions are being proposed as therapeutic adjuncts to facilitate clinical applications in recent decades. They would allow therapists to be emancipated from physically intensive work, and more advanced and interesting rehabilitation training could become accessible to patients with low cost. To date, several rehabilitation devices have been developed or revised to perform bilateral exercises. A new therapy method has been proposed in recent decades, based on the finding that bilateral exercises might promote functional improvement through the reinforcement of corticospinal pathways from the intact hemisphere to the affected arm. However, the safety and reliability of existing bilateral rehabilitation devices is yet to be confirmed, and the cost will be high if the device is expected to provide true bilateral exercises in three-dimensional (3D) space. The goal of this dissertation is therefore to develop a new bilateral rehabilitation system for upper-limb rehabilitation, aiming for safe, stable, and interesting bilateral training for stroke survivors. Some biological data can be measured objectively and in a timely fashion through the proposed bilateral rehabilitation system, which would be analysed to assess recovery stages and further explore bilateral recovery processes. The proposed bilateral rehabilitation system has been developed, validated and utilised in four studies discussed in this thesis. The first one is to develop an industrial robot-based bilateral rehabilitation device for upper limbs including hardware and software, which should be capable of providing a safe and stable training environment. The second one is to develop an intelligent bilateral training subsystem, which is expected to provide true robot-assisted bilateral exercises to cover different recovery stages. The third one is to explore muscle activation patterns during different bilateral training based on the proposed bilateral device and training protocols. These muscle activation patterns can be used as the foundation for the development of bilateral devices, training protocols, and assessment criteria. The last one is to develop a biological signal-based evaluator, which can select different training protocols according to personal biological data. So far, a total of 18 healthy participants have been recruited for evaluating the proposed bilateral rehabilitation system (13 participants tested the device and 5participants tested the evaluator). The results of the experiment show that the proposed bilateral rehabilitation device is safe and stable, the training protocols are reliable and interesting, and the evaluator is objective and time-saving. Meanwhile, the muscle activation patterns gathered using the system are informative, which can help in understanding the cooperation mechanism of each pair of muscles under different bilateral training conditions, and further evaluate the effectiveness of robot-assisted bilateral training. In summary, the developed bilateral rehabilitation system and a comprehensive and systematic study from this work has demonstrated that such a system has the potential for clinical applications with safe, reliable, stimulating, and informative bilateral exercises.