Tracking Targets with Bearing Data from a Single Sensor
Abstract
This report examines performance limits (Cramer-Rao bounds) on tracking a maneuvering target using bearing measurements from a single sensor on a maneuvering platform. An approximation to the Cramer-Rao bound for estimating the location, Velocity, and acceleration of a constant acceleration target with a prior distribution of the target's velocity and acceleration is derived for the case where the target and the sensor are coplanar. The bound is computed for members of a two-parameter family of sensor trajectories, and optimal sensor trajectories within this two-parameter family are identified from contour plots of the bound vs the two parameters. The optimal trajectory in most cases is a weave around the line of sight to the target, with a period which is proportional to the observation time allotted for the measurement. The bound of performance is not in general very sensitive to either the sensor's or the target's motion, or to mismatch between the two, except that the period of the sensor's weave pattern influences both the time at which good estimates become available and the variance of the estimates after a given time interval. Cramer-Rao bounds, Sensor-target geometry, Optimal angle-only tracking, Target tracking, Target acquisition, Lower bounds, Iterative least squares, Extended Kalman filter.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 13, 1989
- Accession Number
- ADA205422
Entities
People
- W. H. Gilson
Organizations
- Massachusetts Institute of Technology