Fundamental Limitations of Optical Trackers.
Abstract
A problem is considered to determine the tracking capabilities of an estimator, applied in optical sensing. The estimator tracks the centroid of a one-dimensional Gaussian-shaped intensity based on time-space point process measurements. The centroid assumes to move dynamically as a First Order Gauss-Markov process. Filter performance is described by steady-state upper and lower bounds on mean-square-error (MSE) which are evaluated as a function of two physically motivated parameters: average number of photons detected in a coherence time of centroid dynamics and mean square of the beam width. The parameters establish regions of operations where upper and lower bounds converge to the actual MSE. Results from the Monte-Carlo simulation demonstrates the bounds usefulness. Noise measurements from dark current or background radiation are included in the simulation. Results show that the filter is very sensitive to these measurements, resulting in very poor tracking. Ad hoc methods of filter tuning and residual monitoring are employed to improve tracking performance; results indicate that filter performance can be improved substantially through residual monitoring. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1978
- Accession Number
- ADA064737
Entities
People
- John M. Santiago Jr
Organizations
- Air Force Institute of Technology