Adaptable DSP Functions for Dynamically Reconfigurable Communication Systems

摘要

The advent of Dynamically Reconfigurable Systems (DRS) is providing new opportunities and challenges for Intellectual Property (IP) designers. It opens up possibilities for novel implementation of well established algorithms and functions but at the same time demands new methodologies and ingenuity. Digital signal processing (DSP) is one area where Dynamic Reconfiguration (DR) can be used to come up with flexible, efficient, low power and cost effective implementations for modern multimedia and wireless communication systems. The objective of this work is to come up with new designs of perfected signal processing techniques that exploit partial dynamic reconfiguration to achieve higher efficiency in modern system implementations. The candidate algorithms, a Linear Predictive Coding (LPC) synthesis filter and an adaptive Recursive Least Squares Lattice (RLSL) filter are tailored for hardware efficiency keeping in view how to reap further benefits from dynamic reconfiguration of the function parameters. The implementations show how dynamic reconfiguration can be used to achieve a better area and performance trade-off in dynamic operating environments. Besides the improvements regarding implementation efficiency of the given DSP algorithms, the achieved profiles on algorithm implementation and application support our related work on distributed buffers schemes and DR architecture model presented in [5] and [6].