Hardware-Implemented Calibration Algorithms for Antenna Arrays

Abstract

An adaptive calibration technique based on the least mean squares (LMS) algorithm is presented which does not rely on integrated calibration networks and is therefore suitable for e.g. array fed reflector antenna systems. Using hardware-implemented algorithms, makes the calibration real-time capable and independent of external computation. Possibilities to enhance the LMS by step-size controlling techniques are introduced and compared by simulation, regarding their adaption speed. The work focuses on the transition of the presented algorithms towards digital hardware, especially FPGAs. The implementation feasibility of the proposed techniques is assessed and timing considerations as well as resource-effective realization are discussed. Suitable algorithms are selected and their implementation on an FPGA is described. Differences to simulation are outlined and optimization through pipelining is explained. The performance of the realized algorithms is evaluated on a Ka-band receiver system. Figures of merit are the adaption speed and the resulting SNR improvement. Step-size control offers slight performance improvements over basic LMS with fixed step-sizes. This comes at the expense of higher resource demand and implementation effort.