The problem of optimum data quantization for memoryless signal-detection systems operating in m-dependent noise environments is considered. The case where quantizer breakpoints are fixed or predetermined is considered first, and, for this case, existing results for general (unquantized) memoryless detection are modified to yield necessary and sufficient conditions for quantizer optimization in terms of asymptotic efficiency. Necessary conditions are also established for the optimum (asymptotically efficient) selection of quantizer breakpoints, and expressions are presented for the comparison of quantizer-detector performance on the basis of asymptotic relative efficiency. (Author)
Discussion(0)
No comments yet. Be the first to comment.