Research Of Intelligent Multi-functional Nursing Wheelchair Bed

As an innovative nursing device in the medical care industry,the multi-functional nursing bed plays an important role in the nursing of the elderly especially disabled and patients.With the progress of society,the aging trend of population is becoming increasingly serious.In today’s society,the nursing methods of the elderly and patients are becoming more demanding for the function of the nursing bed.Considering to satisfy the demand,in this paper,we research on designing a kind of multi-functional nursing wheelchair bed based on speech recognition.Using SolidWorks software as a design platform,we design the three-dimensional modeling of the wheelchair bed including back support,rollover and raising and curving mechanisms by adopting the modularization design method;with the analysis of motion and mechanics,some graphs about the motion angle,angular velocity,angular acceleration and support force of main mechanisms are drawn.It’s proved that the motions of different mechanisms have rationality and the design requirements are satisfied.Based on ANSYS software,the static analysis using the ultimate force loading on the main mechanism is carried out,the rationality of each mechanism is checked from two aspects of strength and stiffness.The stability of two modes of multi-function nursing wheelchair bed system is analyzed and the requirements are met.Based on wavelet transform theory,an improved continuous threshold function algorithm is proposed.The algorithm and the classic hard and soft-threshold algorithms are used to reduce the noise under speech signals.Compared with them,The improved one has a better result for reducing the noise.By simulating on Matlab software,speech Signal-to-Nosie Ratio(SNR)is higher than another two and Root-Mean-Square Error(RMSE)is lower than them.Two threshold endpoint detection algorithm is applied to extract effective speech signals.According to the multi-function nursing wheelchair bed,the host computer software for speech recognition is developed,and specific vocabularies ofthe nursing wheelchair bed functions is processed with the wavelet threshold algorithm and the endpoint detection algorithm.Calling back interface functions from IFLYTEK realizes speech recognition and the function training templates are established by using the XML file technology to improve the accuracy of speech recognition.In relatively quiet and noisy environment,we have voice control experiments on software system that we bulid.The accuracy and stability of speech recognition are tested in system and the influence of environment and specific phrases on recognition accuracy is analyzed.The results showed that the recognition accuracy in the quiet environment basically reached90% and above,the recognition accuracy in the relatively noisy environment reached 85%or more.In a word,the recognition accuracy satisfies the design requirements,voice control are achieved in multi-function wheelchair bed.