Gridlocking and correlation methods and arrangements

Details Gridlocking and correlation methods and arrangements Gridlocking and correlation methods and arrangements

2002-03-08

2004-10-12

Font, Frank G. (Department: 2877)

Optics: measuring and testing

Contour plotting

C382S278000

Reexamination Certificate

active

06803997

ABSTRACT:

FIELD OF INVENTION
The present invention is in the field of tracking objects. More particularly, the present invention provides a method, apparatus, system, and machine-readable medium to correlate objects and gridlock sensors with images based upon track data from sensors such as radars, global positioning systems, laser target designators, seismic sensors, and the like.
BACKGROUND
Correlating objects and gridlocking radars can potentially provide a more accurate depiction of a theater, or model of an area of interest, because the theater is based on tracks from multiple radars rather than one radar. Tracks include data sensed about an object or target by a sensor such as radars, global positioning systems, laser target designators, seismic sensors, and the like, and the data can include the positions and velocities of planes, ships, troops, or other targets. Correlating objects, or correlation, is a process of comparing tracks from different radars to determine which tracks are duplicate tracks. The goal of correlation is to reduce the number of redundant or duplicate tracks so that a theater accurately depicts the objects present within the area of interest, providing a Single Integrated Picture (SIP). Correlation requires substantial statistical analysis in many applications since each radar tracks the position of the object with respect to location and orientation known with uncertainty, especially in applications wherein one or more radars may change positions over a period of time. A global positioning system and compass system at each radar estimates the location and orientation within a margin of error; however, it can be a relatively crude estimate, especially when the platform is in motion. As a result, correlating objects is typically accomplished on a powerful computer by statistical, associative methods. There have been many such approaches over the years for correlation including simple nearest neighbor algorithms, probabilistic data association filtering, multiple hypothesis testing, and others. The nearest neighbor association compares each track of a sensor with each track of another sensor individually to determine the cost of matching a pair of tracks, or cost of determining that the tracks correspond to the same object. The assignment or the Munkres assignment algorithm, for example, assigns the tracks to pairs based upon the example, assigns the tracks to pairs based upon the least overall costs for the matched pairs and satisfies the sequentially most probable hypothesis approach.
Gridlocking, or sensor registration, is a process of combining unique objects from more than one radar by determining the difference in position and orientation of the radars based upon the matched pairs. The goal of gridlocking is to reduce navigation errors and sensor misalignment errors so that one sensor's track data is accurately transformed into another sensor's coordinate system, e.g. radar-aligned. Gridlocking is necessary in many applications for the same reasons that correlation needs significant statistical analysis, the unknown and misaligned radars. The misalignment of tracks from the radars will cause errors when combining objects from different radars to generate the theater. Gridlocking is a data processing algorithm with a predictor-corrector type architecture, typically accomplished on a powerful computer that compares differences in position and orientation of matched pairs with techniques such as Kalman filtering or a similar weighted least-squares approach to find a bias error vector. This bias error vector is used to compute positional adjustments (PADS) to the tracks reported from a given radar. The coordinate transformation adjusts the tracks of the sensors to match the coordinate system chosen for the theater, one track of each matched pair of tracks is discarded (or are possibly combined), and the remainder of the tracks, including the tracks unique to each of the multiple radars, generate or update the theater. However, a problem involved with gridlocking is that gridlocking one radar with another radar based upon pairs of tracks for the same objects is difficult when a pair of tracks determined to describe the same object is inaccurate. For instance, rough correlations of the tracks include mismatches and each individual track pair within an overlap of the two radars, including mismatched pairs, is statistically compared based upon distances and velocities of the tracks to minimize distance and orientation differences. In addition, some systems attenuate the errors by effectively averaging the statistical distance and orientation differences at the expense of additional computing power and time when multiple, candidate track pairs are provided. Further, individual treatment of tracks pairs or tracks for gridlocking, as well as correlation, gives rise to an increase in computational demand with each additional track, or increase in data, effectively precluding the recalculation of coordinate transformations and matched pairs with prior or historical track data.
The computer balances the desire for improved depictions of the theater and the computation demands. An increase in the data that is collected and processed about objects at a given moment in time can potentially increase the accuracy of the theater. For example, a radar drops targets or objects on occasion so that radar will not transmit a track for the object, however, the object will probably be tracked by at least one radar when the object is within the range of two or more radars. The Cray, or similar, computer performs the loops and nested loops of computation for statistical analyses with low latencies to allow more data to be processed. However, the Cray computer gives rise to a problem of portability and latencies inherent to transmitting data to the computer and from the computer back to the theater rather than moving the computer near the sources of data.


REFERENCES:
patent: 4832447 (1989-05-01), Javidi
patent: 5111515 (1992-05-01), Javidi
patent: 5119443 (1992-06-01), Javidi et al.
patent: 5367579 (1994-11-01), Javidi et al.
patent: 5485312 (1996-01-01), Horner et al.
patent: 5600373 (1997-02-01), Chui et al.
patent: 5699449 (1997-12-01), Javidi
patent: 5794173 (1998-08-01), Schutte
patent: 5841907 (1998-11-01), Javidi et al.
patent: 5903648 (1999-05-01), Javidi
patent: 6002773 (1999-12-01), Javidi
patent: 6021378 (2000-02-01), Reiter et al.
patent: 6104336 (2000-08-01), Curran et al.
patent: 6104345 (2000-08-01), Tweg et al.
patent: 6202033 (2001-03-01), Lange
patent: 6225942 (2001-05-01), Alon
patent: 6233357 (2001-05-01), Li et al.
patent: 6282496 (2001-08-01), Chowdhary
patent: 6289132 (2001-09-01), Goertzen
patent: 6317688 (2001-11-01), Bruckner et al.
patent: 6332030 (2001-12-01), Manjunath et al.
patent: 6343155 (2002-01-01), Chui et al.
Tien-Hsin Chao & Hua-Kuang Liu, Real-time Optical Holographic Tracking of Multiple Objects, Applied Optics, vol. 28, No. 2, Jan. 15, 1989, pp. 226-231.
Eugene Hecht, Adelphi University, Optics, Third Edition, pp. 437-438, 442, 532, 456, 457, 602-604, Addison-Wesley.
Kenneth H. Fielding, Joseph L. Horner, Charles K. Makekau, Optical Fingerprint Identification by Binary Joint Transformation, Optical Engineering, vol. 30, No. 12, Dec. 1991, pp. 1958-1961.
Subhasis Saha, Image Compression—from DCT to Wavelets: A Review, ACM Crossroads Student Magazine, http://www.acm.org/crossroads/xrds6-3/sahaimgcoding.html., pp. 1-14.
H. John Caulfield, Templates for Invention in the Mathematical and Physical Sciences with Applications to Optics, Proceedings of SPIE vol. 4392, 2001, pp. 104-117, SPIE.
Eddy C. Tam, Francis T. S. Yu, Don A. Gregory, Richard D. Juday, Autonomous Real-Time Object Tracking with an Adaptive Joint Transform Correlator, Optical Engineering, vol. 29, No. 4, Apr. 1990, pp. 314-320.
Yu Hen Hu, University of Wisconsin, Madison, Visual Artifact Reduction By Post-Processing, Presented at Carnegie Mellon University, Dept, ECE, Nov. 1, 2001, pp. -69.
Shahriar Mirabbasi, On the Design of Multirate Filter Banks by H ~ Optimization,

Affiliated with

Also associated with

Rating

Gridlocking and correlation methods and arrangements does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Gridlocking and correlation methods and arrangements, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Gridlocking and correlation methods and arrangements will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3316194