An Approach of Constructing Sensitive Metric for State Estimation Algorithm

For the performance evaluation of state estimation algorithm in the field of information fusion, this paper analyzes the shortcoming of conventional evaluation indexes and proposes a constructing method of sensitive index based on optimal subpattern assignment. The sensitivity of this metric can be reasonably distributed to location error and association error by using two controllable parameters. By simulating under two different conditions, it is shown that OSPA distance not only has the same validity and accuracy of performance evaluation as the mean square root error of position in a singletarget environment, but also has higher flexibility and sensitivity in a multitarget environment.