Achieving Real-Time Efficiency for Adaptive Radar Pulse Compression

The adaptive pulse compression (APC) algorithm has been shown to effectively suppress the range sidelobes resulting from large scatterers to the level of the noise thereby unmasking nearby small scatterers and yielding substantial sensitivity improvement over deterministic filtering techniques. However, implementation in a real-time system is limited by the relatively high computational complexity of the algorithm. As such, this paper considers dimensionality reduction techniques as a means to reduce this computational cost in order to enable real-time operation. In general, numerous different methods exist to decompose the minimum mean-square error (MMSE) framework upon which the APC algorithm is based. In this paper dimensionality reduction is achieved either by segregating contiguous blocks of the received signal samples or by decimating the received signal. Each of the two approaches then facilitate the approximation of the original MMSE cost function of APC by the sum of lower dimension MMSE cost functions thus yielding filters that require a lower computational cost to implement. The reduced dimension algorithms, collectively denoted as the fast adaptive pulse compression (FAPC) algorithm, substantially reduce the computational cost of APC while still maintaining much of its performance benefit.