A Review of Data Fusion Techniques

In many cases, researchers use more than one sensor and synthesize their raw data to generate more meaningful information that can be of greater value than single source data. The process of merging multiple data and knowledge from different sources to represent the object into a regular, accurate, useful, meaningful representation is known as data fusion. This article summarizes the state of data fusion and compares relevant techniques. We explain possible data fusion classifications and review the most common fusion methods such as Kalman filter and The Bayesian Methods. Then we evaluate these methods and discuss the advantages and disadvantages of each method. General Terms Multi-sensor fusion, data fusion, Kalman filter, Particle filter, Bayesian methods, Dempster-Shafer.