Targets Existence Decision Based on RBFN and Bayes' Detection Theory
Zong Chengge · Chuangan jishu xuebao · 2007
By making use of chaos characters of sea clutter and Ne ural Network's goodness of non-linear fit,a RBF network for prediction is des igned to decide whether there are targets in radar echo wave or not.The problem of targets detection can be viewed as two-patterns decision through calculatin g forecasted mean-squared error on either of targets-including imput and targe ts-non-including imput.We take Bayes' detection theory,based on minimum fal se ratio,as the rule of decision.On assuming that the transcendent probabiliti es of the two patterns are equal,the crossing point of two patterns' condition al probability density curve can be made the cut-off in this emulation,from wh ich the minimum signal-to-noise ratio for detection is obtained as-13.271 dB.