Simultaneous M-ary Channel Hypothesis Test with Least-Mean-Square Signal Amplitude Estimation

摘要

This paper presents a new strategy for detecting and estimating multiple linearly independent signals immersed in additive random noise. This new approach is a Maximum Log-likelihood Ratio Test (MLRT) and determines the presence of one or more signals through a single filter operating on a received input data vector. In addition, it performs a signal-amplitude estimation when a positive response occurs. This test is a UNiformly Most Powerful Invariant Test and reduces the computationally intensive work to only those situations where important data exists. The estimation technique determines the signal amplitude of all possible input signals in a least-mean-square sense with a variance inversely proportional to the inherent signal-to-noise ratio of each linearly independent library waveform. Computer simulations verify the theory and demonstrate near-optimum performance of the MLRT for small signal libraries. The MLRT is close to optimum for larger signal sets, and is clearly optimum for complex multiwaveform signals.