A Bioelectrical Signals-Based Upper Limb Rehabilitation Robot System

With the improvement of socioeconomic level,people are paying more and more attention to the quality of life,especially to the demand of medical service.The increasing number of paralysed patients has prompted the born of new rehabilitation training and evaluation systems with the help of highly deveoped medical science and technology.The study of biological electrical signals(EEG,EMG,etc.)in rehabilitation medicine research field has opened up a new path for traditional rehabilitation therapies.Brain electrical signals which are generated by the paralysed patients themselves contain their active consciousness,where intentions of the patients can be extracted and interpreted into communication or control signals to assist them to do rehabilitation training,thus the impaired neural system can be restored.The brain-computer interface(BCI)technology developed in recent decades provides a strong technical support for that.In this paper,we present an independent rehabilitation training system for upper limb to help patients achieve active training.In the study we combine BCI technology and robot technology to achieve the design of the whole system.In the system,emotional signals which are easy to control are used to trigger the beginning and ending of the training,while motor imagery brain signals which are valuable to brain neural plasticity are selected to trigger every step move in training.Moreover,a robot system is responsible for concrete enforcements to alleviate mental fatigue of patients,as well as improve the training efficiency and reliability.In the other part of the study,a novel rehabilitation assessment method is put forward.Electromyogrphy(EMG)signals are analyzed and processed to evaluate the patient's rehabilitation condition.EMG signals which are generated by the muscles are different while muscle contracting and relaxting.Different EMG patterns may indicate different musule force.As EMG signals are chaotic and its characteristic of nonlinear is obvious,in this paper we perform an analysis on EMG signals based on the approximate entropy(Ap En),and ceate a two-dimensional characteristic vector with ApEn and electromyographic signal intensity.With the quadratic discriminant analysis(QDA)method,65% classification accuracy was got.Our research provides a new way of training and evaluation method for upper limb in rehabilitation field and beneficial references to the related scholars,which may help in broadening the ideas of research.