Data fusion using expected output membership functions
1999
Multi-sensor systems can improve accuracy, increase detection range, and enhance reliability compared to single sensor systems. The main problems in multi-sensor systems are how to select sensors, model the sensors, and combine the data. This dissertation proposes a new data fusion method based on fuzzy set methods. The expected output membership function (EOMF) method uses the fuzzy input set and the expected fuzzy output. This method uses the intersections of the fuzzy inputs with the expected fuzzy output in order to find relationships between the given inputs and the estimate of the output. The EOMF method creates a fuzzy confidence distance measurement by assessing the "fusability" of the data. The fusability measure is used for finding the best position of the EOMF and the best estimate of the system output. Adaptive methods can help deal with occasional bad measurements and set the EOMF to the proper width. The EOMF method can be used for both homogeneous and heterogeneous sensors, which give redundant, cooperative or complementary information. In addition, the EOMF method is robust in the sense that it can eliminate sensor measurements that are outliers. The EOMF method compares favorably with other methods of data fusion such as the weighted average method. An example from the control of automated vehicles shows the effectiveness of the adaptive EOMF method, compared to the fixed EOMF method and the weighted average method in the presence of Gaussian and impulsive noise. This method can also be applied to nondestructive evaluation (NDE) images from heterogeneous sensors.
Keywords:
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
39
References
0
Citations
NaN
KQI