Transformation to implement computationally efficient IIR decimation filters

A new technique is proposed to transform an IIR filter into a computationally efficient decimation filter. In this technique, the recursive transfer function is transformed and noble identity is invoked to get sample rate reduction. The magnitude and phase response of the original filter are unaltered after the transformation. A higher order IIR transfer function is decomposed into parallel first-order sections and each section is transformed for sample rate reduction. The transformation is computationally efficient since current output can be directly computed from single Mth old output and M inputs processed using polyphase decomposition. Filtering and down sampling are performed in the same stage. Filters are designed and mapped on FPGA. Hardware and computational complexities are reduced and throughput is enhanced.