Title :
Neural networks in frequency bin detection
Author :
Hsieh, Luke S L ; Wood, Sally L. ; Shetty, Ravi R.
Author_Institution :
Dept. of Electr. Eng., Santa Clara Univ., CA, USA
Abstract :
This paper studies the application of neural networks in frequency bin detection. The structure of the neural network, the initial condition of weights, and the creation of the training patterns for this specific application determine the dimension and the convergence behavior of the neural network. After being trained, the neural network can detect the presence of multiple tones from time sampled data, which may contain amplitude errors, phase errors, and additive noise. Remedies for errors are proposed via modifying weights or increasing the dimension of the neural network. Simulation results are given to support the analysis
Keywords :
backpropagation; convergence; error analysis; error compensation; feedforward neural nets; noise; signal detection; backpropagation; convergence; dimension; error analysis; error compensation; feedforward neural networks; frequency bin detection; multiple tone signals; time sampled data; training patterns; weight initialisation; Convergence; Error analysis; Feedforward neural networks; Feedforward systems; Frequency; Intelligent networks; Neural networks; Neurons; Parallel processing; Signal processing algorithms;
Conference_Titel :
WESCON/'93. Conference Record,
Conference_Location :
San Francisco, CA
Print_ISBN :
0-7803-9970-6
DOI :
10.1109/WESCON.1993.488478