Research And Application Of Rehabilitation Training Expert System Of Case-based Reasoning

Rehabilitation training for hemiplegia parts is a very important and critical medical treatment to cerebral apoplexy patients, while clinical rehabilitation method for cerebral apolexy patients heavily depends on a treatment of one-to-one physical therapy which is between a doctor and a patient. However, in that case the training method's efficiency and intensify is difficult to guarantee, and the training effects mostly relies on physician's treatment level. What's more, the lack of quantitative and objective parameter datas between training evaluations and rehabilitation effects is another obstacle to optimize the training parameters for obtaining the best treatment method.Given that situation, this thesis researches and designs a rehabilitation training expert system of case-based reasoning, which performs the rehabilitation training for the patients through controlling a upper limbs rehabilitation robot. By use of a 5-DOF upper limbs rehabilitation robot and related technology to monitor and evaluate the training process objectively, can not only improve pertinence and scientificity of hemiplegia rehabilitation training, but also can survive physicians from the heavy manual labor to make better rehabilitation program, so as to improve the efficiency of rehabilitation further. In the analysis and design of rehabilitation training expert system, this thesis adopts the method of CBR to solve the problem of knowledge acquisition and knowledge representation. This method can expresse expert knowledge in the form of cases, and can be used in system inference process.This thesis firstly introduces the hemiplegia rehabilitation medicine theory and exoskeletons 5-DOF upper limbs rehabilitation robot system's main components. Secondly, it expounds the theory of expert system and CBR method, mainly discusses the basic structure and function of expert system, the basic composition and working principle of CBR. Then, combined with rehabilitation cases and object-oriented representation method to represent the rehabilitation cases. It adopts the k-nearest neighbor method to retrieve similar cases. And comes up with proportion interval method to process the data of rehabilitation cases,so that these data could be at between 0 and 1. Meanwhile, it makes a research on the calculation of attribute weights. Thirdly, this thesis uses Visual c++6.0 as front-desk system development tool, builds the knowledge base in the form of case base, uses Microsoft Access 2003 to design and create the database and tables. It builds the case base of system by establishing table relationship,and then builds a rehabilitation training expert system based on case reasoning to make a control scheme for the upper limbs rehabilitation robot. Finally, this thesis tests the expert system and the test results show that this system is effective.