Position determination from modern radar systems-new theoretical results and their practical aspects

Several position estimation algorithms have been elaborated till now under assumptions that can be called classical ones. However, these algorithms may be generally ineffective when they are applied in modern radar systems. The problem of the object position estimation has been formulated under general assumptions corresponding to the real measurement conditions that exist (occur) in modern radar systems. The presence of unknown systematic measurement error (sensor misalignment errors, bias) and of navigation or geodetic errors (sensor position uncertainty) are taken into consideration. A general observation model is introduced. This model, called the extended observation model, takes into account the above-mentioned errors. An appropriate unbiased estimator of the object position has been obtained using the generalized least squares criterion. The estimator is presented in a compact form using a rectangular matrix pseudoinverse (Moore-Penrose generalized inverse). The observation model and estimator are general, and all known observation models and position estimators can be considered as particular cases of the observation model and estimator obtained.