Font Size: a A A

Design Of Low-Power Smart Wireless Sensor For Acoustic Fingerprint Recognition

Posted on:2022-03-21Degree:MasterType:Thesis
Country:ChinaCandidate:H LvFull Text:PDF
GTID:2492306608495034Subject:Electronic Science and Technology
Abstract/Summary:PDF Full Text Request
As an important equipment of the power system,the real-time monitoring of its operation status and fault early warning play a very crucial role in guaranteeing the quality of power supply of the power system.As one of the core construction platforms of ubiquitous power IOT,the intelligent transformer fault online detection system based on wireless sensor network technology aims to achieve remote monitoring of transformer status and real-time warning of faults,etc.As one of the key technologies in the implementation of the fault detection system,the wireless acoustic sensor node chip micro-power technology directly affects the system application life and maintenance cost.It has become a research hotspot in both academic and industrial circles.This paper intends to closely focus on the micro-power technology of wireless acoustic sensor chip,focusing on the chip architecture,operating mode and circuit design to carry out micro-power technology research to solve the problem of sensor low power consumption.The main work of the paper contains the following parts:the MFCC-SVM transformer fault audio identification model is established through MATLAB software,low-power design techniques such as logic optimization and resource reuse are adopted,and the hardware logic design of the Mel Frequency Cepstral Coefficient(MFCC)extraction algorithm is completed using the Verilog HDL hardware description language;RISC-V architecture was used to design hardware acceleration extension instructions for the support vector machine(SVM)classification algorithm based on the E203 core,and the hardware and software design of the extension instructions was completed;the sensor multi-voltage domain and multi-frequency domain chip design and implementation techniques were studied,and the power and clock management modules were designed and implemented according to the system operation mode,including the design of programmable multi-output linear regulators(LDO)and phase-locked loops(PLL);the software and hardware test platform for sensor architecture was built,including the debugging of the RISC-V GCC tool chain and the selection of FPGAs.Finally,the prototype verification of the smart sensor digital logic circuit based on the test platform shows that the sensor can effectively complete the detection of transformer faults with an accuracy of 98.131%and a recognition response time of 0.7562ms.
Keywords/Search Tags:RISC-Ⅴ, Mel Frequency Cepstral Coefficients(MFCC), Support Vector Machine(SVM), Acoustic fingerprint
PDF Full Text Request
Related items