Gesture Recognation And Muscle Characteristics Evaluation Exploration Base On Single And Dual Frequency Ultrasound Transducers

There are some deficiencies for the existing human-machine interfaces(HMI),which seriously affect the clinical application.Compared with other physiological signals,ultrasound-based HMI has advantages of high spatial resolution,ability to detect the information of deep muscles and harmlessness.However,the equipment of B-mode ultrasound in medical field is bulky and expensive,which is not wearable.In addition,there is always a tradeoff between resolution and penetration depth.Therefore,a novel HMI based on single and dual frequency single-element ultrasound(A-mode ultrasound)transducers has been proposed and applied to dexterous gesture recognition and muscle characteristics evaluation.First of all,this paper analyses the situation of mainstream HMIs and then focuses on advantages and related problems of ultrasound-based HMI,thus determines main contents of this paper.Secondly,this paper studies structure design and simulation of single frequency A-mode ultrasound transducers.The frequency of transducer is determined as 5 MHz.The 1-3 composite piezoelectric with thickness of 0.34 mm and diameter of 6 mm is selected.The normalized thickness and acoustic impedance of matching layer are 0.2-0.3 times of wavelength and 0.2-0.35 times of 1-3 composite,respectively.Then,performances of customized and commercial transducers are compared and the superiority of customized transducers in detecting human tissue has been proved.On this basis,structures of horizontal stacked and annular array dual frequency A-mode ultrasound transducers and parameters simulation have been studied in this paper.After this,the simulation is verified by comparing acoustic pressure and spectrum at the same position.Comparing performances of single with dual frequency transducers,the potential of dual frequency transducers has been demonstrated.Then,this paper conducts the dexterous gesture recognition with single frequency,horizontal stacked and annular array dual frequency ultrasound transducers.The worst accuracy is still 94.79%.Meanwhile,compared with single frequency ultrasound transducers using whether LDA(p = 0.0039)or SVM(p = 0.0015)classifier,the accuracy has an obvious increase.The results show the advantage of dual frequency ultrasound transducers in gesture recognition.Finally,this paper studies the relationship between muscle force and muscle thickness measured by dual frequency ultrasound transducers.In addition,this paper also studies the muscle fatigue.In the experiment of muscle force,the muscle thickness gradually increases with the force increasing,while the muscle thickness gradually decreases with the force decreasing.Through fitting the normalized force and thickness,it is obviously found that "hysteresis" occurs between muscle force ascending stage and descending stage.In the muscle fatigue experiment,it can be found that the muscle thickness gradually increases during fatigue,and the rising speed of muscle thickness is faster at the first stage and slower at the second stage.All of the above can confirm the effectiveness of our proposed program in this paper.