Effect of linear and nonlinear signal processing on signal statistics
A. V. Balakrishnan · Journal of Research of the National Bureau of Standards Section D Radio Science · 1964
The term " Sig na l P rocessing" is i nterpreted as a n operat ion on th e sig nal t h at is deli berate (as di st ing ui sh cd fr o m the u nav oidable) a nd a ri sin g fr o m :(a) Opt imi za tion: a n o peratio n on th e data t o opt imi ze extract ion or d et ection a~ in r ad ar or co mm un icatio n or s imil ar a ppli cation s.(b) Rou t ine d a ta ha nd lin g : o peratio ns s uch as sampli ng a nd /or q ua ntitizing, sca le cha nges, etc.(c) Ad ap t ive tec hn iques : oper a t io ns t h at a re ch a racteri st ic of ad a p t ive syste ms w here no a prio ri kn owledge of signals and/or syste m p a rameters m a y be a vaila ble and a selflearni ng p roced ure is n ecessary .The main in ter est is in t he eff ec t of t hese operatio ns o n th e sig na ls; more specifi call y, if we r epresent t he p rocessor as a blac k-box, t he sig nal being its " i np ut ," \I'e st ud y the stat istics of t he " o utput" sig na l.We exam ine cer t ain as pec ts of t hese proble ms of r ece nt sig ni fi cance to r adi o phys ics .