Design of digital filters for communication systems

Digital filters are finding extensive applications in modern communication systems because of their inherent advantages over their analog counterparts. Raised-cosine type of digital filters are used frequently in pulse shaping because of their ability to minimize the error probability caused by the Intersymbol Interference (ISI). In this paper we discuss the design of raised-cosine digital filters required to meet some specific design objectives imposed by the desire to obtain minimum ISI and architectural simplicity. Two optimal methods, viz, the Remez exchange method and the Linear Programming method to design digital filters are compared. When the filter specifications satisfy certain conditions called "half-band" conditions, the computationally efficient Remez exchange method is preferable to the Linear Programing technique. The report also discusses the design of optimally partitioned linear phase transmit and receive filters for minimum ISI in the overall filter. Filter design examples are included.