Adaptive lattice filter implementations on pipelined multiprocessor architectures

Three pipelined multiprocessor implementations of adaptive lattice filters are examined. A performance analysis is done for each multiprocessor system, and expressions for approximate computation time and speedup are derived. Each system is shown to be flexible with respect to the different algorithms and the different filter sizes it can implement.<
>