This thesis deals with multisensor fusion in the presence of systematic errors inthe context of target tracking. A typical target tracking problem consists of multiplesensors producing position measurements of multiple targets, for instanceaircraft. The goal is to establish tracks on all targets that are observed by the sensors.A track usually consists of an id, such as a track number, target position,and target velocity. The systematic errors are modeled as measurement biases. Ifunaccounted for these biases may lead to inaccurate estimates of the target state(position and in particular velocity) and a single target may appear as several targetsif the biases are large enough (ghost tracks). Furthermore, if all sensors arebiased it is challenging to find an unbiased estimate of target state with respectto a coordinate system independent of the sensors.In this thesis the sensors are radars producing measurements in 3D. The systematicerrors (biases) are called alignment bias, location bias and sensor bias.The first two are related to sensor deployment, as they describe errors in orientation(misalignment) and sensor placement (location). The sensor bias addresseserrors caused by sensor imperfections. These biases are estimated relative to asensor independent coordinate system and relative to a sensor of reference (mastersensor). A novel distinction is made in this context, where a universal biasestimator (UBE) is used relative to sensor independent coordinates, while an absolutebias estimator (ABE) is used relative to a master sensor. The estimabilityof the biases is investigated using a novel estimability index, which quantifieswhether a bias can be estimated more accurately with the available measurements.The estimability index is based on the Cramer-Rao Lower Bound. The study of estimability is used to determine a multisensor-multitarget scenariowhere several bias estimators are compared with respect to performanceusing a Monte Carlo simulation. The simulation includes alignment, locationand sensor biases, and all sensors are affected. The estimators are evaluated insensor independent coordinates and master sensor coordinates. Two Kalman Filter(KF) based estimators are used as references. A lower bound is representedby a KF where the bias values are known, while an upper bound is representedby a KF where the measurement noise is increased to reflect the biases present.The alignment, location and sensor biases contain three elements each, to a totalof nine bias values to estimate per sensor. The UBE performs well (belowthe upper bound) in sensor independent coordinates when one of the sensor biasvalues are removed from the simulation, estimating eight bias values per sensor.Performance is close to the lower bound when the location bias only is removed,yielding six bias values per sensor. In master sensor coordinates the ABE hasthe best performance. However a simplified version has almost identical performance.It is called the Relative Bias Estimator (RBE), and it neglects the biasesof the master sensor. This is a popular assumption in the literature, and this studyconfirms that this simplification should be preferred in an implementation.Possible extensions of this work are explored. First curved target motion is exploredby letting the target move at constant altitude above the Earth. The curvatureof the trajectory results in increased bias estimability. However, observingthis curvature requires observing the target for a long time with high accuracy.This is challenging in practice, and therefore this path was not explored further.Second, extending the application to Air Traffic Control (ATC) is considered. Atairports radars typically produce 2D measurements, so to extend the developed3D bias estimators it is necessary to incorporate altitude measurements from theaircraft Mode C transponders with these 2D measurements. The altitude measurementsare quantized and received with a coarse resolution which may havea negative impact on bias estimator performance since the vertical velocity estimatebecomes unstable. Several estimators are developed to estimate altitudeand vertical velocity, and these are tested on real measurement data for a performancecomparison. The main contribution is the use of the Interacting MultipleModel (IMM) and Unscented Kalman Filter (UKF) based estimators on quantizedreal world measurements. The UKF produces the best performance forlong term altitude predictions, meaning that its vertical velocity estimate is themost stable.
展开▼
