Design of Radial Basis Neural Network Filter for Pulse Compression and Sidelobe Suppression in High Resolution Radar
- 1 May 2006
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 21555745,p. 1-4
Abstract
Two novel techniques using Radial Basis Neural Networks are presented to deal with pulse compression and sidelobe suppression. Optimal Binary Codes (OBC) are taken as input vectors and the network is trained off-line with desired output response. Signal to Sidelobe Ratio (SSR) and Integrated Sidelobe Level (ISL) are considered as evaluation parameters for the proposed techniques. One technique uses Correlated Radial Basis Neural Network Sidelobe Suppression (CRBNSS) Filter algorithm and the other uses Mismatched Radial Basis Neural Network Sidelobe Suppression (MRBNSS) Filter algorithm. In this paper, the proposed filters are designed at optimum SPREAD of the network in presence of range resolution and Doppler shift with additive noise to achieve high SSR and low ISL. The simulation results show that both the filters result in superior performance in range resolution ability, Doppler tolerance and noise rejection compared to the conventional filters. Ambiguity diagrams are presented for better analysis and comparison.Keywords
This publication has 8 references indexed in Scilit:
- Cross-correlation of long binary signals with longer mismatched filtersIEE Proceedings - Radar, Sonar and Navigation, 2005
- A Neural Fuzzy Network Approach to Radar Pulse CompressionIEEE Geoscience and Remote Sensing Letters, 2004
- Performance characteristics of pulse radar detection using neural networks2002
- Pulse compression for weather radar2002
- Improving performance in pulse radar detection using neural networksIEEE Transactions on Aerospace and Electronic Systems, 1995
- Design of low-sidelobe pulse compression waveformsElectronics Letters, 1994
- A neural network approach to pulse radar detectionIEEE Transactions on Aerospace and Electronic Systems, 1993
- The merit factor of long low autocorrelation binary sequences (Corresp.)IEEE Transactions on Information Theory, 1982