Frequency domain system identification and it's application in jamming

Abstract

Identification is a powerful technique used to build accurate models of system from noisy data. The Frequency Domain System Identification utilizes specialized tools for identifying linear dynamic multiple input/single-output (MISO) systems from time responses or measurements of the system's frequency response. Frequency domain methods supports continuous-time modeling, which can be a powerful and highly accurate complement to the more commonly used discrete-time methods. The methods described here can be applied to problems such as the modeling of electronic, mechanical, and acoustical systems. A brief intuitive introduction to system identification using adaptive filters has been provided. Adaptive filtering has been broadly divided into seven broad categories .The first part describes linear and non-linear filtering and gives an overview of various kinds of estimation techniques. The second part describes Weiner filter and its applications in real world. The third part introduces system modeling like stochastic and stationary models. In the fourth part we describe the least mean square algorithm and it’s error performance. The fifth part gives an overview of variants of LMS like sign LMS, block LMS, normalized LMS. The sixth part introduces transform domain adaptive filtering and algorithms used like DCT/DFT LMS. The last part is an implementation of system identification in design of a jam resistant receiver.