A simulated evaluating system for multi-sensor data fusion algorithms
Yansheng Lu, Yue Pei, Binbin Qu · 2004
Data fusion is the process to synthetically calculate information coming from several sensors, which helps reduce possible mistakes or uncertainty of apperceiving targets in the information process. A general performance evaluation platform is introduced for data fusion algorithms, which can be used for the distributed multi-target tracing process. The improved insert-value method was implemented to simulate the original three-dimension data and a disturb model which considers the sensor performance and uncertain factors was applied. Furthermore, we put forward a criterion to assess the synthetic performance. Efficiency and validity were substantiated by the simulated examination. For testing one or several data fusion algorithms, which are available, or being investigated, the system can supply a practical appraising method.