SoC System Design Of Emg Decoding And Control Of Prosthesis Based On FPGA

The SoC design method based on Intellectual IP is proposed,to satisfy the requirements for lower power consumption,lower cost and shorter development cycles.However,with the increase of design scale and complexity,the construction cycle and cost arise caused by repeated chips production.Thus,a method of a SoC prototype design verification based on FPGA has been proposed.FPGA has the high flexibility and offers the large amount of IP core authorization,the SoC design method based on FPGA develop fast and become the main method.This paper introduces the design method of SoC system based on FPGA in detail according to decoding sEMG.The generation mechanism and physiological characteristics of EMG signals are described in this paper,and the sources and characteristics of interference part in EMG signal acquisition are also analyzed,and a complete optimization solution for interference signals is proposes.The EMG signal is magnified by the hardware circuit and filtered by hardware and software,followed by the extraction of the signal with high signal-to-noise ratio and integrity.The collected useful EMG signal is analyzed by Native Bayes Classifier(NBC)after the pretreatment and feature extraction.A new hardware structure for EMG decoding is proposed.The decoding module is achieved by using logic circuit,and the computing performance is evaluated.The resources and characteristics of the hardware platform are introduced in detail in the construction of FPGA-based SoC system.The system hardware platform and embedded software architecture are constructed respectively.The overall design of the EMG decoding SoC system is completed by calling the decoding IP.The development cycle is significantly reduced under the hardware and software co-design.Meanwhile,the flexible reconstruction features can meet different requirements.The paper designs an experimental paradigm of the mechanical arm manipulating object motion to verify the system.Finally,the mechanical arm accomplishes specified task under remote control after off-line training of the classifier through the combination of hardware and software simulation.The experimental results show that the hardware decoding of the EMG has faster operation speed,the accuracy of the decoding rate is up to 92% which can assure the mechanical arm completes the specified task successfully,and also verifies that FPGA-based SoC makes the design more flexible and efficient.