Quickest detection of a time-varying change in distribution

A practical algorithm for quickest detection of time-varying arbitrary one-parameter changes in a sequence of independent random variables is developed. The amplitude of the parameter need not to be known. This model can be applied to the problem of coherent detection of sampled sinusoidal signals of known frequency, but unknown phase and amplitude. The tests are designed according to a maximum allowable false alarm rate. Expressions that predict algorithm performance, in terms of average detection time are obtained. Simulation results show the scheme has improved performance over E.S. Page´s (1954) quickest-detection procedure in the detection of sampled sinusoids of known frequency (and unknown amplitude and phase) in white Gaussian noise