ATI's Multi-Target Tracking and Multi-Sensor Data Fusion course

The objective of this course is to introduce engineers, scientists, managers, and military operations personnel to the fields of radar tracking, data fusion and to the key technologies which are available today for application to this field. The course is designed to be rigorous where appropriate, while remaining accessible to students without a specific scientific background in this field. The course will start from the fundamentals and move to advanced concepts. This course will identify and characterize the principal components of typical tracking systems. A variety of techniques for addressing different aspects of the tracking data fusion problem will be described. For example, different techniques are required for the assimilation of "time-late" data than those used for "real-time" data. Real world examples of data fusion systems used by both the Navy and the Marines will be presented and discussed. This course will also use specific illustrative examples to show the tradeoffs and systems issues between application of different techniques. 

What you will learn:

• State Estimation Techniques – Kalman Filter, constant-gain filters.
• Non-linear filtering – When is it needed? Extended Kalman Filter.
• Techniques for angle-only tracking.
• Tracking algorithms, their advantages and limitations, including:
  - Nearest Neighbor
  - Probabilistic Data Association
  - Multiple Hypothesis Tracking
  - Interactive Multiple Model (IMM)
  
• How to handle maneuvering targets.
• Track initiation – recursive and batch approaches.
• Architectures for sensor fusion.
• Sensor alignment – Why do we need it and how do we do it?
• Attribute Fusion, including Bayesian methods, Dempster-Shafer, Fuzzy Logic.

Course Outline:

 

1. Introduction. Basic concepts & definitions. Target motion models, measurement models, coordinate systems.

    

2. The Kalman Filter. State estimation – least squares and Kalman filtering.

    

3. Other Linear Filters. Constant-gain and table look-up filters. Comparative performance.

    

4. Non-Linear Filters. When are they necessary? Extended Kalman Filter, Unscented Filter, Particle Filters.

    

5. Angle-Only Tracking. EKF, pseudo-linear, modified polar coordinates.

    

6. Maneuvering Targets: Adaptive Techniques. Noise models, maneuver detection, adaptive processes.

    

7. Maneuvering Targets: Multiple Model Approaches. Non-switching and switching models. Interacting Multiple Model. IMM design. Examples.

    

8. Single Target Correlation & Association. Correlation processing. Nearest neighbor, assignment algorithms.

    

9. Track Initiation, Confirmation & Deletion. M/N and sequential tests for confirmation. Track deletion criteria.

    

10. Using Measured Range Rate (Doppler). Implementation of constant-gain, Kalman and EKF using measured range rate.

     

11. Multitarget Correlation & Association. Extended nearest neighbor, optimal & sub-optimal assignment.

     

12. Probabilistic Data Association. Examples and implementation issues.

     

13. Multiple Hypothesis Approaches. Examples. Hypothesis merging and pruning.

     

14. Coordinate Conversions. Conversions between local systems and from local to global. Compensation for sensor motion.

     

15. Multiple Sensors. JDL Data Fusion Model. Levels of fusion.

     

16. Data Fusion Architectures. Fusion architectures. Report-to-Track and Track-to-Track. Associated Measurement Reports

     

17. Fusion of Data From Multiple Radars. Comparison of Report-to-Track and Track-to-Track processes. Colocated and non-colocated sensors.

     

18. Fusion of Data From Multiple Angle-Only Sensors. Correlation. Triangulation techniques for non-colocated.

     

19. Fusion of Data From Radar and Angle-Only Sensor. Correlation techniques. Differences between collocated and non-colocated sensors.

     

20. Sensor Alignment. Types of alignment problems. Impact of biases. Techniques.

     

21. Fusion of Target Type and Attribute Data. Bayesian, Dempster-Shafer, Fuzzy Logic. Impact of attribute data on correlation.

     

22. Performance Metrics. Quantizing system performance. Measures of track accuracy, continuity, initiation time, etc.