Unified Registration Model for Both Stationary and Mobile 3D Radar Alignment

J. Geolo. Geo-inf. Geo-int.

Progri³

2018

Self Cite

Registration is the act or process of simultaneously entering information about radar offset biases and attitude biases of Inertial Navigation System (INS) in a common reference frame known as radar registration frame. The variables that describe the registration are defined as radar offset biases and attitude biases. The mathematical relationship that shows the formal relation among the registration variables or radar offset biases and attitude biases is defined as the registration equation(s). For mobile 2-, 3-D radar networks, in order to improve the accuracy of the registration equations, the earth-centered earth-fixed (ECEF) coordinate system is used as the intermediate frame to get to the registration frame. Since a 2-D radar model has no comprehensive 3-D information of the target when establishing the registration equations, three different methods are discussed, each method differently and uniquely pinpoints to a reference point. The latter instead of the true target coordinates is used to establish the registration equations. The errors of the reference points are unavoidable and will be embodied in the estimation errors of all the biases. The mobile 2-D radar network registration in ECEF frame can also be included in this methodology. Simulation results show that all three methods can be applied well for registration, and for a certain condition, one of them outperforms the others.

Section: True Target Coordinates In Enu Framementioning

confidence: 99%

“…In order to make the registration equations observable, the equivalent azimuth and elevation biases are modeled as a constant plus a noise, respectively [6], where the noise arises…”

Section: The State Equationsmentioning

confidence: 99%

Section: The State Equationsmentioning

confidence: 99%

See 1 more Smart Citation

Mobile 2-, 3-D Radar Network Error Registration Model

J. Geolo. Geo-inf. Geo-int.

Progri³

2018

Self Cite

“…It is assumed that both radars are synchronized with the same sampling intervals T = 5 s. Chen et al, 2013;Kay, 1993;Progri, 2011).…”

Section: C H E N a N D O T H E R S Vol 67mentioning

confidence: 99%

Analysis of Mobile 3-D Radar Error Registration when Radar Sways with Platform

et al. 2013

J. Navigation

For mobile radars installed on a gyro-stabilised platform (GSP) that can steadily follow an East-North-Up (ENU) frame, attitude biases (ABs) of the platform and offset biases (OBs) of the radar are linear dependent variables. Therefore ABs and OBs are unobservable in the linearized registration equations; however, when combining them as new variables, the system becomes observable, and this model has been called the unified registration model (URM). Unlike GSP mobile radars, un-stabilised GSP (or UGSP) mobile radars are installed on the platform directly and rotate with the platform simultaneously. For UGSP, it is testified that both types of biases are independent and observable because the time-varying attitude angles (AAs) 1 of the platform are included in the registration equations, which destroy the dependencies of both kinds of biases and lead us to propose a completely different linearized registration model-the All Augmented Model (AAM). AAM employs all OBs and ABs in the state vector and a Kalman filter (KF) to produce their estimates. Numerical simulation results show that the estimated performance of AAM is close to the Cramér-Rao lower bound (CRLB) and that the Root Mean Square Errors (RMSEs) of the rectified measurements by using AAM are more than 500 m smaller than by URM in all directions. K E Y

“…Intuitively, just as in the previous discussion in [2]- [5], the relational expressions of TTCs between MF and ENU could be obtained by using trigonometric functions according to the geometry shown in Figure 2. In fact, this method does not work because concise TTCs in ENU cannot be written out in this way.…”

Section: Introductionmentioning

confidence: 97%

Quaternionic analysis of the influences of systematic biases on radar data processing

J. Geolo. Geo-inf. Geo-int.

Progri³

2015

Self Cite

For single radar tracking, Kalman filter (KF) can be used only when radar measurements have Gaussian white noises. However, when systematic biases exist, application of KF for single radar tracking is not trivial. Furthermore, in the existing radar network, the association (both measurement-to-measurement and measurement-to-track) problem and the systematic biases removal are mixed together, which further complicates radar network data fusion. The key to solving these problems is to find and apply a set of mathematical tools (quaternionic analysis) on the biased measurements to directly enable the application of KF. This paper provides a detailed discussion of quaternionic analysis that can be applied to solve this problem. Measurement frame (MF) is proposed where target states contain radar systematic biases. It is proved that for stationary and first kind of mobile radar (installed on gyro-stabilized platform) target kinematic model (TKM) in MF has the same form with that in East-North-Up (ENU) frame. And KF can be used directly for the biased measurements to obtain the biased target track in the sense of minimum mean square error (MMSE). However, for the second kind of mobile radars (rigidly connected and sways with platform), a first-order approximation yields large errors therefore they (or second kind of mobile radars) cannot be applied well when attitude biases are greater than 1°. This situation will be investigated and discussed further in future publications.