A high speed complex adaptive filter for an asymmetric wireless LAN using a new quantized polynomial representation

This paper introduces a new number representation, the Quantized Polynomial Representation (QPR), which is used for building FIR filters where some quantization errors can be tolerated. The technique is based on the previously published Modulus Replication Residue Number System (MRRNS) but considerable savings are possible if polynomial quantization can be tolerated. The QPR can be used as a vehicle for Quadratic Residue Number System (QRNS) mapping of complex data, and the main computational architecture can be built with independent finite ring computational channels. We demonstrate this new technique on an asymmetrical Gigabit wireless LAN where the adaptive filter computes with complex arithmetic. We demonstrate area savings of up to 28% and power savings of up to 50%.