Assessment of data fusion systems part 2: Sensor network architectures

In the previous paper, we described an assessment methodology that provides us with a more well-rounded interpretation of data fusion (DF) system performance based on complexities of the ground truth and sensor reports. The methodology in that paper yields a more accurate performance indication of a data fusion system as compared to a methodology based solely on output performance. However, we still need to incorporate other significant input factors to further address the complexity of scenarios in order to fortify our assessment methodology. A key area that was not incorporated in our previous paper is communications among nodes within a sensor network. Communications are integral to the functioning of a sensor network DF system and can drastically affect the quality of its outputs, especially in harsh hostile environments. In this paper, an additional series of complexity metrics is incorporated with regards to the architectures and communication factors of data fusing sensor networks. These metrics are incorporated with the metrics in the previous paper to obtain an improved assessment index of the fusion system performance.