| Sensing technology is the foundation of information science. Sensor technology is one of the important footstones of the modern information technology, which has almost permeated through everywhere of science and technology and national economy. Mankind feels the outside information through visual, olfactory, auditory and tactile sensory and human brain processes these information to control the action of human body. This is the basic instinct of mankind to understand and rebuild the world. When it has to say the computer is the extension of human brain, the sensor is the extension of human facial features.In this paper, details of initial, construction and parameters determine for a fast-converging discrete binary wavelet neural networks (WNN) has been introduced. This kind of WNN is used in sensors non-liner error correction and a simulation result is given. Compare to other neural networks trained by stochastic gradient method, theory of discrete binary wavelet transform speeds up the training procedure of fast-converging WNN and makes this WNN avoid converging to a local maximum point. This fast version of wavelet neural network is also not complicate in building. It solved the problem of slow convergence of dilation and translation parameters which caused by differentiating wavelets. This paper analyzed abilities of approximation and predication with an example of approximation and prediction highly non-liner function with limits sample point. After that, this paper uses this WNN to back modeling thermocouple correction table and learn the correction curve when the temperature of reference point of thermocouple other than zero, and gave a result of it.At last, this paper discusses the topic in FPGA implementation of WNN. This detail discuss include compute number system, implementation of transmit function and implementation of scale and translation. And Join Matlab and Modelsim to simulate the function of this FPGA implementation model. |
PDF Full Text Request