A taxonomy of voting schemes for data fusion and dependable computation

Voting is an important operation in the fusion of data originating from diverse sources and in the realization of ultrareliable systems based on multiple computation channels. Voting involves the derivation of an output data object from a collection of n input data objects, as prescribed by the requirements and constraints of a voting algorithm. The objects voted on can be quite complex in terms of content and, explicitly specified or implicit, structure. Regardless of implementation details (e.g., whether realized in hardware, software, or hybrid schemes) and object space properties, voting algorithms can be classified according to how they view the input and output data objects and how they handle the votes (weights) at input and output. A 16-class (binary 4-cube) categorization results from dichotomizing each of the above voter features. This categorization leads to an abstraction that helps in the study of voting algorithms with regard to the dependability level for the outputs and the speed at which they are obtained; viz, the quality and efficiency of the algorithms. The taxonomy is broad enough to cover, and detailed enough to distinguish among, a wide variety of commonly used voting algorithms in data fusion and dependable computation. It also provides insight into the relationships of various voting schemes and facilitates comparison and fine-tuning of such algorithms.