Data merging
Theory and practice
This course is for learners who want to discover the techniques of data merging and to understand their benefit.
Following an examination of the tools and theoretical resources, learners will be able to apply these techniques to concrete cases of information synthesis and automatic recognition.
The course lasts 5 days (30 hours) and includes:
- 18 hours of theory-based lectures
- 12 hours of examples of processing
The course takes a theory-based approach to explore methods of data merging.
Many examples, distributed during the lectures, will help participants to master the application of these methods.
They will learn how to evaluate and design a multi-sensor system.
Course level: Advanced
This course is intended for engineers and decision-makers.
A good knowledge in statistics would be greatly beneficial.
Jean-François GRANDIN:
Engineer at the Technical Department of Electronic Warfare Systems at THALES Airborne Systems.
- Data merging
- Introduction to the course
- Combination of uncertain information
- Theories, probabilities with fuzzy measures
- Combination operators
- Robust merging
- Example: multi-sensor passive situation
- Statistical estimation and decision-making
- Statistical estimation:
- Estimator, criterion
- Bias, variance, CRAMER-RAO terminal
- Exponential laws
- Maximum likelihood (ML) recursive
- Conditional expectation
- Post maximum estimation (PME)
- Statistical decision-making:
- Motivations and definitions
- Problem with simple hypotheses
- Bayesian and NEWMAN-PEARSON criteria
- Ratio of likelihood and thresholds
- Decision-making problems with composite hypotheses
- UMP test - statistical invariance
- Examples of applications
- Objective designation for a two-mode missile
- Location using distributed goniometers
- Plot/track merging
- Statistical estimation:
- Multi-sensor tracking
- Tracking Loop - Track Management
- Modeling: Models of targets, observation
- MLE - Kalman Filter - Bayesian Filter
- Multi-Model Filtering (IMM) - Merging Tracks
- Examples of application:
- Performance analysis from simulations
- Air/air and air/surface applications
- Multi-sensor tracking
- Combination processing, information used
- Modeling: probabilistic, combinatorial approach
- Classical Algorithms: PMHT - MHT - Bayesian Approaches - SD - Assignment
- Examples of combination processing
- Image merging
- Image processing
- Feature extraction and recognition
- Summary
- Recap of the course - Development of a multi-sensor system - Criteria for choosing the constituent elements
- Final panel discussion
Scheduled in French:
PARIS: 14 to 18 June 2021
For the English realization, please, consult us.
