Research Of Training Program And Evaluating System For 5-DOF Upper Limb Rehabilitation Robot

With the aging of the world have become increasingly prominent, and the number of stroke disease are increasing, thousands of people are deprived of one or more exercise capacity. The motor deficit seriously have affected their quality of life, and brought a heavy burden on society and family. At present, the clinical treatment of hemiplegia rehabilitation is mainly limited to rehabilitation physician, who trained and evaluated patients one-to-one by hands. Rehabilitation physician’s work is intensive. The treatment effect mainly depends on physicians’skill level, clinical experience, and training equipment, etc. The clinical treatment is difficult to achieve high strength, targeted and repetitive rehabilitation training. Thus, this dissertation introduced the robotics and intelligent control theory and methods into the field of clinical rehabilitation medicine in order to improve the rehabilitation level of stroke patients, which had a great significance on stroke patients, their family and our society.This dissertation studied the issue of rehabilitation training program and evaluation for 5-DOF upper limb rehabilitation robot system. In this dissertation, the main works were as follows:(1) For the large deviations of weight coefficient in the Delphi method, the weighted Delphi-Fuzzy evaluation method was proposed. The proposed method can reflect patients’ condition fully. And then, the evaluation model of motor function was established to achieve comprehensive evaluation for different stages of the function rehabilitation of patient’s upper limb.(2) A new automatic case acquisition approach based on hierarchical clustering was presented. When the similar case cannot find in the case base, cluster merge the case, whose knowledge intensity greater than or equal to the minimum knowledge intensity, increase the cases number in the case base and obtain the similar case, and then store the case into the case base.(3) A method of case revise based on different attribute matrix was proposed. When the most similar case does not meet the conditions of case reuse, if the most similar case will be directly reused, it cannot get a satisfactory solution. Therefore, used the deviation between the most similar case and the rehabilitation patients’ condition to revise the solution, and then obtained optimal rehabilitation training program.(4) In the construction method of membership function for interval type-2 fuzzy sets, the shape of footprint of uncertainty (FOU) is unknown. Aiming at this issue, an approach for the interval of upper and low membership was put forward. The proposed method designed a new survey question to collect the interval data of upper and lower membership degree, after data preprocessing, determine the membership function of interval type-2 fuzzy sets effectively.(5) In the Enhanced Karnik-Mendel algorithm of interval type-2 fuzzy sets, the initial value of switch point is uncertain, and the computation rate is low. For these problems, a novel EKM algorithm with bidirectional search was developed. The proposed method made the initial value is uniquely determined, and can search up and down, and then improved the computation rate effectively.(6) In the process of drawing up a rehabilitation training program, there is diversity, uncertainty and complexity of patients’ condition. So the hierarchical clustering-interval type-2 method was developed, combing with the case-based reasoning (CBR) of hierarchical clustering and interval type-2 fuzzy reasoning (IT2FR) with bidirectional search. Experimental analysis showed that the method can make a rehabilitation training program.(7) Designed and developed the rehabilitation training program and evaluating system for 5-DOF upper limb robot. And the comparison of the patients’ clinical data can indicate that the system’s evaluation result and rehabilitation training program are consistent with the clinical rehabilitation experts’ results. It means that the system can assist rehabilitation experts to evaluate stroke patients, and make a rehabilitation training program effectively.