Application oftheMLPDA toBistatic Sonar

Approximately adecade agothemaximumlike-viewpoints: that is, totrack using datafusion. Multi-sensor lihood probabilistic dataassociation (MLPDA)tracking ar- processing bestows manyadvantages, including different and chitecture wasproposed; itwasfound, viasimulation, tobe perhaps offsetting Doppler blind velocities, uncertainty re- averyeffective (perhaps theonly) waytotrack verylow- duction viacovariance intersection (particularly whenrange observable contacts - downto7-8dB(post signal process-accuracy isfarbetter thanbearing accuracy, orvice-versa) ing), meaning that inalarge field ofviewagiven truecon- (7), corroboration ofdetections, andofcourse theobvious tact might onlyrankconsistently amongthetopfive orsixof oneofincreased aggregate target return (improved fused de- those obtained. TheMLPDA isbased onmaximizing statisti- tection probability) (11). callikelihood according toaprecise model. Itsapplication to real multistatic data (which isunlikely tofit MLPDA assump- Covariance intersection isanotional toolthatclarifies the tions) isbothatest ofits effectiveness andanopportunity to benefits ofestimation viameasurements whosecovariance explore newwaystoimprove it.We shall seeinwhatfol- matrices areeccentric andarenotaligned. Researchers have lows: bistatic operation; automatic registration viathe"directshownthat this complementarity oferrors isencouraged via blast"; theuseoftheSavage amplitude feature; false-alarm operation inamultistatic mode(2). A standard bistatic active density modification; andasmartgrid-search. sonarisoneinwhichtheping emanates fromonevessel and isobserved atadifferent location. Thisdivision ofduties can