System for correlating object reports utilizing connectionist architecture
First Claim
1. A system for correlating object reports in a first set of multi-dimensional data with object reports in a second set of related multi-dimensional data, said multi-dimensional data including x coordinate position, y coordinate position and velocity, said first set comprising tracking data vectors of predicted object reports at a given time and said second set comprising vectors of sensed object reports at said given time, said system comprising:
- host computer for receiving and storing said vectors of predicted object reports and said vectors of sensed object reports;
a plurality of input means for accepting said vectors of sensed object reports in said second set of data from said host computer;
a set of processing elements each producing a correlation output signal that is a difference calculation between said vectors of sensed object reports and said vectors of predicted object reports;
a set of interconnections coupling each of said input means with each of said processing elements, said interconnections each having a fixed strength that is a function of factors of a polynomial for calculating said difference multiplied by said value of said tracking vectors of predicted object reports, said weighted connections to each processing elements also being proportional to one of said tracking vectors in said first set of multi-dimensional data, said processing element multiplying said vectors of said sensed object reports by said interconnection strength; and
controller mean for comparing the correlation output signal from said processing elements for each vector of a predicted object report and each vector of a sensed object report and for assigning each of said vectors of object reports in said second set to one of said vectors of object reports in said first set based on the maximum correlation signal produced by a said processing elements, wherein said controller means further comprises a host computer for determining if correlation signals are greater than a minimum threshold before performing said assignment, said host computer also storing said maximum correlation signals and determining if a given correlation produced by an object report from said second set in a given processing element is greater than previous maximum correlation signals produced in said processing element by a different object report in said first set.
1 Assignment
0 Petitions
Accused Products
Abstract
A correlation processor (22) for solving large correlation problems involving multi-dimensional data (12, 16, 20). The system correlates vectors in a first set with vectors in a second set. The system includes a plurality of input devices (24) for accepting the entities (16) in the second set of data (14). Also, a set of processing elements (26) are included for producing a correlation output signal that is a first predetermined function of its inputs. A set of variable strength interconnections (28) couple each of the input devices (24) to each of the processing elements (26) through a connection having a strength that is a second predetermined function of the first set of multi-dimensional data (12), wherein the weighted connection (28) to each processing element (26) is proportional to one of the entities in the first set of multi-dimensional data. A controller (48) is provided for assigning each of the entities (16) in the second set (14) to one of the entities in the first set based on a comparison of the correlation signals from the processing element (26). The system requires minimal algorithm development, minimal software development and minimal preprocessing. It can be implemented using current hardware technology at a reasonable cost to provide real-time solution to multi-dimensional correlation problems.
32 Citations
9 Claims
-
1. A system for correlating object reports in a first set of multi-dimensional data with object reports in a second set of related multi-dimensional data, said multi-dimensional data including x coordinate position, y coordinate position and velocity, said first set comprising tracking data vectors of predicted object reports at a given time and said second set comprising vectors of sensed object reports at said given time, said system comprising:
-
host computer for receiving and storing said vectors of predicted object reports and said vectors of sensed object reports; a plurality of input means for accepting said vectors of sensed object reports in said second set of data from said host computer; a set of processing elements each producing a correlation output signal that is a difference calculation between said vectors of sensed object reports and said vectors of predicted object reports; a set of interconnections coupling each of said input means with each of said processing elements, said interconnections each having a fixed strength that is a function of factors of a polynomial for calculating said difference multiplied by said value of said tracking vectors of predicted object reports, said weighted connections to each processing elements also being proportional to one of said tracking vectors in said first set of multi-dimensional data, said processing element multiplying said vectors of said sensed object reports by said interconnection strength; and controller mean for comparing the correlation output signal from said processing elements for each vector of a predicted object report and each vector of a sensed object report and for assigning each of said vectors of object reports in said second set to one of said vectors of object reports in said first set based on the maximum correlation signal produced by a said processing elements, wherein said controller means further comprises a host computer for determining if correlation signals are greater than a minimum threshold before performing said assignment, said host computer also storing said maximum correlation signals and determining if a given correlation produced by an object report from said second set in a given processing element is greater than previous maximum correlation signals produced in said processing element by a different object report in said first set.
-
-
2. A method for correlating entities in a first set of multi-dimensional data with entities in a second set of related multi-dimensional data, said entities in the first set comprising tracking data vectors of predicted object reports at a given time and said entities in the second set comprising vectors of sensed object reports at said given time, said method comprising the steps of:
-
receiving said vectors of sensed object reports in said second set of data in a plurality of input units; coupling each of said input units to a set of processing elements through a connection having a fixed strength that is a function of factors in a polynomial;
using said polynomial to calculate the difference between the vectors of sensed object reports and the vectors of predicted object reports, multiplied by said vectors of predicted object reports, said weighted connection also being proportional to one of said tracking data vectors in said first set of multi-dimensional data;generating correlation output signals in a plurality of processing elements that are a function of the sensed object report vectors multiplied by said weighted connection; comparing the correlation output signals from said processing elements for each predicted object report and each sensed object report; and assigning each of said object reports in said second set to one of said object reports in said first set based on said comparison of said correlation signals from said processing elements. - View Dependent Claims (3, 4, 5)
-
-
6. A method for correlating entities in a first set of multi-dimensional data with entities in a second set of related multi-dimensional data, said entities in the first set comprising tracking data vectors of predicted object reports at a given time and said entities in the second set comprising vectors of sensed object reports at said given time, said method comprising the steps of:
-
receiving said vectors of sensed object reports in said second set of data in a plurality of input units; coupling each of said input units to a set of processing elements through a connection having a fixed strength that is a function of factors in a polynomial for calculating the difference between the vectors of sensed object reports and the vectors of predicted object reports, multiplied by said vectors of predicted object reports, said weighted connection also being proportional to one of said tracking data vectors in said first set of multi-dimensional data; generating correlation output signals in a plurality of processing elements that are a function of the sensed object report vectors multiplied by said weighted connection; comparing the correlation output signals from said processing elements for each predicted object report and each sensed object report; and assigning each of said object reports in said second set to one of said object reports in said first set based on said comparison of said correlation signals from said processing elements, wherein said step of assigning further comprises the step of determining which of said correlation signals are greater than a minimum threshold before performing said assignment, the step of storing said maximal correlation signals and also the step of determining if a given correlation produced by an object report from said second set in a given processing element is greater than previous maximum correlation signals produced in said processing element by a different object report in said first set.
-
-
7. A system for correlating object reports in a first set of multi-dimensional data with object reports in a second set of related multi-dimensional data, said multi-dimensional data including x coordinate position, y coordinate position and velocity, said first set comprising tracking data vectors of predicted object reports at a given time and said second set comprising vectors of sensed object reports at said given time, said system comprising:
-
host computer for receiving and storing said vectors of predicted object reports and said vectors of sensed object reports; a plurality of input means for accepting said vectors of sensed object reports in said second set of data from said host computer; a set of processing elements each producing a correlation output signal that is a difference calculation between said vectors of sensed object reports and said vectors of predicted object reports; a set of interconnections coupling each of said input means with each of said processing elements, said interconnections each having a fixed strength that is a function of factors of a polynomial, said polynomial configured to calculate said difference multiplied by said value of said tracking vectors of predicted object reports, said weighted connections to each processing element also being proportional to one of said tracking vectors in said first set of multi-dimensional data, said processing element multiplying said vectors of said sensed object reports by said interconnection strength; and controller means for comparing the correlation output signal from said processing elements for each vector of a predicted object report and each vector of a sensed object report and for assigning each of said vectors of object reports in said second set to one of said vectors of object reports in said first set based on the maximum correlation signal produced by said processing elements. - View Dependent Claims (8, 9)
-
Specification