Research On Rehabilitation Expert System For Upper-Limb Rehabilitation Robot

In today’s China, stroke survivors are often treated in the department of neurology for only14days before they are left without proper follow-up care out of the hospital. Many stroke survivors suffer from "disuse syndrome" being short of adequate rehabilitation treatment, while many others have to undergo "misuse syndrome" as a result of inappropriate treatment. Therefore, it is very important to provide stroke survivors sufficient and correct rehabilitation treatment as soon as possible. It is the concern of our whole society to offer enough medical resource to hemiplegia victims.One solution is to apply robot to rehabilitation to bring about reliable self-help treatment. As a well-developed artificial intelligence technique, expert system is broadly used in areas like medical diagnose and control engineering. Even its application has not been seen very often in rehabilitation robot, it shows signs of suitability and this should be a good opportunity. This thesis concentrates on the development of rehabilitation expert system. The main work is as follows:(1) To introduce relevant medical theory on rehabilitation and discuss the necessity and effectiveness of motor training. Evaluating and integrating the overall hardware and software on the robot to facilitate further research on how to provide control strategy;(2) To confirm that employ expert system as the control strategy provider is both feasible and suitable. Rehabilitation expert system can serve as assistive diagnose tools for users, which can be patients and doctors, to give suggestions, as well as a commander for the robot;(3) Designed a knowledge base based on Brunnstrom evaluation scale, in which the parameters carrying support level are consulted from a professional rehabilitation physician. I devised two mode, which is safe mode and quick mode as an attempt to provide safe and easy-used software. Actually, safety is the constant consideration throughout the development process and it is mainly achieved by credible knowledge acquisition, favorable knowledge representation and well-designed inference method, while the efficiency is supposed to rise from more intelligent algorithm;(4) Conflict resolution strategy of inference engine of expert system is examined and a new method to carry out the resolution efficiently is proposed. In order to deal with the dilemma of choosing among qualified rules, I design a secondary inference engine for the major one which can provide rules to resolve conflicts among competing candidates.