Performance analysis of auto-regressive UWB synthesis algorithm for coherent sparse multi-band radars

Auto-regressive spectral gap filling algorithms have been demonstrated in conventional multiband radars, exploiting multiple radars at different incoherent frequency bands to synthesize a larger bandwidth and to significantly increase the range resolution. These algorithms require heavy computations for recovering the phase coherence among bands. Recently, the use of photonics has been demonstrated enabling multiband coherent radars. This coherence allows reducing the computational cost of auto-regressive algorithms by approximately 40%, thus enabling their application in realtime target identification. This manuscript presents a numerical as well as experimental investigation of both the effectiveness and the working limits of a modified autoregressive spectral gap filling algorithm for coherent multiband radar system. The accurate reconstruction of the radar cross section has been demonstrated for a target having two scatterers with a mutual distance of 500 mm. The two scatterer's identification was possible by utilizing merely 2.2% of the entire interpolated bandwidth (two 82 MHz subbands in Sand X-band). In the case of using the 15% of the total bandwidth, the error in range accuracy was ±0.04 m for experimental results. Note that without interpolation, a single sub-band alone would reach an accuracy of mere 1.83 m in the best considered (noise-free) case. The accuracy in range resolution and the comparison with data fusion technique has been analysed for varying fractional bandwidth.