Problem characterization in tracking/fusion algorithm evaluation

The performance of a tracking/fusion algorithm depends very much on the complexity of the problem. This paper presents an approach for evaluating tracking/fusion algorithms that considers the difficulty of the problem. Evaluation is performed by characterizing the performance of the basic functions of prediction and association. The problem complexity is summarized by means of context metrics. Two context metrics for characterizing prediction and association difficulty are normalized target mobility and normalized target density. These metrics should be presented along with the performance metrics. The context metrics also support more efficient generation of input data for performance evaluation. Simple tests for evaluating basic tracking algorithm functions are presented.