Implementation of GSC based subarray adaptive LMS algorithm using Xilinx FPGA

Large scale fully adaptive systems are difficult to implement on FPGA because of low latency, high speed data transfer, and computational requirements. A solution is proposed in previous correspondence where randomly sampled adaptive systems are employed to distribute complex problem into blocks of simple subarray modules. In this contribution we analyze generalized sidelobe canceller (GSC) based LMS subarray adaptive algorithm for FPGA implementation in complex domain. A complex adaptive algorithm requires twice as much device area as compared to real adaptive systems. GSC based LMS adaptive algorithms are evaluated using varying number of data sizes and antenna elements. A twelve channel adaptive beamformer with 18 bit resolution fits into almost one fourth of a Xilinx Virtex4 XC4VSX55 device. This means one Virtex4 device can accommodate both the sum and difference subarray channels for estimating target azimuthal and elevation polarization.