Abstract:Only a few publications exist at present on heterogeneous track-to-track fusion (T2TF). A common limitation of the current work on heterogeneous T2TF is that the cross covariance due to common process noise cannot be computed. This is due to the fact that two local trackers use different dynamic models, and hence, it is difficult to account for the common process noise. We consider a heterogeneous T2TF problem in three dimension (3-D) using a passive infrared search and track (IRST) sensor and an active air moving target indicator (AMTI) radar with the nearly constant velocity motion of the target. The active AMTI tracker uses the Cartesian state vector with 3-D position and velocity, and the dynamic model is linear. A passive IRST tracker commonly uses modified spherical coordinates (MSCs) for the state vector, where the dynamic model is nonlinear. In this formulation, the common process noise is explicitly modeled in both dynamic models. Therefore, it is possible to take into account the common process noise. We use the cubature Kalman filter (CKF) in both trackers due to its numerical stability and improved state estimation accuracy over existing nonlinear filters. The passive tracker uses a range-parameterized MSC-based CKF, and the active tracker uses a Cartesian CKF. We perform T2TF using the information filter (IF), where each local tracker sends its information matrix and the corresponding information state estimate to the fusion center. The IF handles the common process noise in an approximate way. Results from Monte Carlo simulations show that the accuracy of the proposed IF-based T2TF is close to that of the centralized fusion with varying levels of process noise and communication data rate.

A High-Level Track Fusion Scheme for Circular Quantities

Harmonic Mean Density Fusion in Distributed Tracking: Performance and Comparison

Track Fusion with Incomplete Covariance Information

An Unscented Kalm an Filter-Based MultisensorTrack Fusion Algorithm

Two Kinds of Sensor Data Fusion in Target Tracking

On Pooling-Based Track Fusion Strategies : Harmonic Mean Density

A Robust Fusion Algorithm For Multi-Sensor Tracking

Track-to-track data fusion for Unmanned Traffic Management System

Distributed Track-to-track Fusion for Non-Linear Systems with Gaussian Mixture Noise

Selective Track Fusion.

Target Tracking Based On A Distributed Neural Network Filtering Fusion

Heterogeneous Track-to-Track Fusion in 3-D Using IRST Sensor and Air MTI Radar

Recursive LMMSE Fusion for Tracking with Non-Colocated Multi-Radar Measurements

A Measurement Fusion Algorithm with Dissimilar Sensors in Case of Arbitrary Correlated Noises

Track Fusion in Multi-Sensor Systems without Knowledge of Noise Covariance

A Robust Fusion Algorithm Based on Multi-Sensor Tracking

Adaptive track association in distributed multi-sensor track fusion system

A Data Fusion Approach for Track Monitoring from Multiple In-Service Trains

Track Association and Fusion Based on Information Demand Analysis

Distributed Multitarget Tracking Via Heterogeneous Fusion with 3D Radars and Infrared Sensors.

Track Fusion in the Presence of Sensor Biases