VLSI Design and Implementation of Least Mean Square Algorithm in Electronic Weighing System

• Main Authors: Shih-Ming Lin, 林世明
• Other Authors: Yeong-Kang Lai
• Format: Others
• Language: en_US
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/37994736403742389152

碩士 === 國立中興大學 === 電機工程學系 === 92 === The weighing of articles is an essential part in the modern life. We usually need to know the weight of the articles, such as food, product components, Pharmacology (medicine weight), chemical agents and so on. In the field of weighing, the object is put on the weighing system with the sensor. The sensor is an uncontrollable measuring component. The weight of object is obtained by weighing and filtering. In the process of weighing in electronic scales, the signals coming from load cell are disturbed by many kinds of noises, such as the air perturbation, the vehicles engine undulation, the surface vibration, the temperature change, the electronic circuit noise, and other uncontrollable factors. These noises will lower the accuracy and the speed of the electronic scale. In order to reduce noises and to increase the authenticity of the signal, it is very important to design an effective digital filter. With the improvement of filtering, we can increase the speed and the accuracy of weighing. We can eliminate all kinds of noise by utilizing digital filters. The digital filters have high stability and the parameter of them can be modulated easily. The filter procedure may be used on any control loop. We have the demand of high performance filtering, but there are only few researches done. In this thesis, we propose a digital filter with high efficiency, and we improve the speed and the accuracy of the electronic scale. We propose an effective filter model and verify the digital filter. The environment of the weighing changes frequently. Therefore, the parameters of weighing model and the filter change frequently. After the analysis and the simulation of digital filtering, we design the proposed FIR architecture with low hardware overhead and high speed. Based on the analysis and implementation of the proposed FIR filter architecture, the stability and the resolution of the electronic scale can be enhanced easily.