Note on Sensor Resource Allocations: Higher Rate or Better Measurements?

When tracking maneuvering targets and the need to improve the tracking arises or designing a sensor tracking system, one is often faced with the choice of increasing the measurement accuracy or rate. The answer to this question is found by assessing the impact on error in the filtered state estimates and the one-step predicted state estimates. In this paper, the tracking of maneuvering targets with a nearly constant velocity (NCV)Kalman filter is considered and the maximum mean squared error (MMSE)is utilized to study the impacts of doubling the measurement accuracy or rate. For each measurement case and the maximum acceleration of the maneuvering target, the process noise variances that minimize the MMSE in the filtered position and the one-step predicted position are used to assess the impacts of doubling the measurement accuracy or rate. The analysis shows that doubling the measurement accuracy gives the greater reduction in MMSE in filtered position, while doubling the measurement rate gives the greater reduction in the MMSE in the one-step predicted position. Selection of the process noise variance that minimizes the MMSE in the one-step predicted position estimates is also new in this paper. Monte Carlo simulation results are given for tracking a maneuvering target with an NCV Kalman filter to verify the findings and an Interacting Multiple Model (IMM)estimator tracking a maneuvering target to assess the generally applicability of the findings.