Panigrahi, N and Tripathy, S (2010) Application of Soft Computing Techniques to RADAR Pulse Compression. BTech thesis.
Soft Computing is a term associated with fields characterized by the use of inexact solutions to computationally-hard tasks for which an exact solution cannot be derived in polynomial time. Almost contrary to conventional (Hard) computing, it is tolerant of imprecision, uncertainty, partial truth, and approximation to achieve tractability, robustness and low solution cost. Effectively, it resembles the Human Mind. The Soft Computing Techniques used in this project work are Adaptive Filter Algorithms and Artificial Neural Networks.
An adaptive filter is a filter that self-adjusts its transfer function according to an optimizing algorithm. The adaptive filter algorithms used in this project work are the LMS algorithm, the RLS algorithm, and a slight variation of RLS, the Modified RLS algorithm.
An Artificial Neural Network (ANN) is a mathematical model or computational model that tries to simulate the structure and/or functional aspects of biological neural networks. It consists of an interconnected group of artificial neurons and processes information using a connectionist approach to computation. Several models have been designed to realize an ANN. In this project, Multi-Layer Perceptron (MLP) Network is used. The algorithm used for modeling such a network is Back-Propagation Algorithm (BPA).
Through this project, there has been analyzed a possibility for using the Adaptive Filter Algorithms to determine optimum Matched Filter Coefficients and effectively designing Multi-Layer Perceptron Networks with adequate weight and bias parameters for RADAR Pulse Compression. Barker Codes are taken as system inputs for Radar Pulse Compression. In case of Adaptive Filters, a convergence rate analysis has also been performed for System Identification and in case of ANN, Function Approximation using a 1-2-1 neural network has also been dealt with. A comparison of the adaptive filter algorithms has been performed on the basis of Peak Sidelobe Ratio (PSR). Finally, SSRs are obtained using MLPs of varying neurons and hidden layers and are then compared under several criteria like Noise Performance and Doppler Tolerance.
|Item Type:||Thesis (BTech)|
|Uncontrolled Keywords:||Adaptive, Artificial Neural Network, Matched Filter, System Identification Barker Codes, LMS, RLS, MLP, Function Approximation, Back Propagation Algorithm, PSR, SSR, SNR, Doppler Shift|
|Subjects:||Engineering and Technology > Electronics and Communication Engineering > Soft Computing|
|Divisions:||Engineering and Technology > Department of Electronics and Communication Engineering|
|Deposited By:||Nishant Panigrahi|
|Deposited On:||13 May 2010 15:23|
|Last Modified:||13 May 2010 15:23|
|Supervisor(s):||Sahoo, A K|
Repository Staff Only: item control page