Research On Intelligent Sickbed Motion Controller Based On Brain-computer Interface

Millions of people in the world are sick in bed for a long time because of a variety of reasons.They can't independently turn over,sit up and so on.The care of these patients requires a lot of manpower and resources.Most of the existing electric nursing beds on the market rely on manual remote controls as man-machine interfaces,and disabled elderly people cannot use them independently.The brain-computer interface can monitor the patient's motion intention by monitoring the patient's EEG signal,so that the patient can independently control the movement of the bed,which not only reduces the workload of the nursing staff,but also improves the patient's sense of dignity.Therefore,this project developed a smart bed motion controller based on brain-computer interface.This article uses a single-channel EEG sensor to control the movement of the bed.At present,the multi-channel brain-computer interface commonly used at home and abroad requires the user to wear an electrode cap,which is very inconvenient.In this study,a single-channel EEG sensor is conveniently used.The recording electrode is located at the Fpl of the forehead to the left,and the reference electrode is located at the left earlobe.In terms of algorithms,blink information and information of closed eyes on the changes in the rhythm of the brain electrical energy(especially changes in the alpha rhythm)is used.Since the bed has 8 actions,it is difficult to directly classify the single channel EEG signals.This paper innovatively uses the binary tree to change the 8 classification into 3 binary classifications,which improves the classification accuracy.When the subject uses the brain-computer interface,it is necessary to put a screen in front of the subject to display a block diagram of the 8-action binary tree.The method of switching the idle state of the bed and the control state of the bed is to blink quickly and continuously.When the subject performs the binary classification,the method of selecting an option is to close the eyes to relax.The process of processing the EEG signal is to first remove the power frequency interference by filtering,and then remove the EOG by the empirical mode decomposition and the principal component analysis algorithm.The energy of each rhythm and its derived parameters are selected as features,and they are classified by the support vector machine algorithm.In this paper,10 research subjects were selected for online classification experiments,and 20 sub-categories were selected respectively,with a classification accuracy of 91.5%.Three subjects were recruited for online real-time control of smart beds,achieving a predetermined goal with an average accuracy of 92%.The main innovations of this paper are as follows:one is to use the single-channel EEG sensor to control the eight movements of the bed;the other is to change the 8 classification into three binary classifications through a binary tree,and then use the blink information and the EEG rhythm transformation information to implement user's intent recognition...