IPC분류정보
국가/구분 |
United States(US) Patent
등록
|
국제특허분류(IPC7판) |
|
출원번호 |
UP-0165779
(2008-07-01)
|
등록번호 |
US-7692573
(2010-05-20)
|
발명자
/ 주소 |
|
출원인 / 주소 |
- The United States of America as represented by the Secretary of the Navy
|
대리인 / 주소 |
|
인용정보 |
피인용 횟수 :
37 인용 특허 :
2 |
초록
▼
A system and methods enable efficient data association of input sensor data with uniquely identified candidate targets. The methods may use information provided by a target status history database, a target geolocation history database, and a target technical characteristics database, as well as dat
A system and methods enable efficient data association of input sensor data with uniquely identified candidate targets. The methods may use information provided by a target status history database, a target geolocation history database, and a target technical characteristics database, as well as data processing procedures provided by an algorithm rules database. The algorithm rules database provides procedures for generating target classification and identification information for input sensor data, for matching target classification and identification information obtained from input sensor data with information provided by the target technical characteristics database to generate an initial set of consistent, uniquely identified candidate targets, for estimating the minimal required speed of advance for each candidate target, for calculating weights and corresponding data association probabilities for the initial set of candidate targets with the input sensor data, and for selecting a final set of uniquely identified candidate targets with their data association probabilities.
대표청구항
▼
What is claimed is: 1. A computer readable storage medium having a method encoded thereon, the method represented by computer readable programming code, executed by a computer to perform the method comprising the steps of: classifying input sensor data by comparing one or more measured target attri
What is claimed is: 1. A computer readable storage medium having a method encoded thereon, the method represented by computer readable programming code, executed by a computer to perform the method comprising the steps of: classifying input sensor data by comparing one or more measured target attributes with one or more known attributes of a finite set of uniquely identified targets; generating an initial set of one or more candidate targets based upon a comparison of the classified input sensor data with the known attributes of the finite set of uniquely identified targets; calculating a minimum required speed for each candidate target to travel between a geolocation history location and a location provided by the input sensor data, the minimum required speed subject to both obstacle avoidance and status transition time constraints; assigning a statistical weight to each candidate target based upon its calculated minimum required speed; calculating a data association probability of each candidate target with the classified input sensor data; and generating a final set of candidate targets by selecting the candidate targets from the initial set of candidate targets having calculated data association probabilities that exceed a predetermined threshold value. 2. The computer readable storage medium of claim 1, wherein the step of generating an initial set of one or more candidate targets includes selecting, from the finite set of uniquely identified targets, a subset of uniquely identified targets having attributes consistent with the classified input sensor data. 3. The computer readable storage medium of claim 1, wherein the step of calculating the minimum required speed for each candidate target includes the steps of: determining a first obstacle avoidance distance, wherein the first obstacle avoidance distance is the distance between a location specified by the input sensor data and a known location of the candidate target at a time value just before a time value of the input sensor data; determining a second obstacle avoidance distance, wherein the second obstacle avoidance distance is the distance between the location specified by the input sensor data and a known location of the candidate target at a time value just after the time value of the input sensor data; determining a first time interval, wherein the first time interval is the difference between the time value of the input sensor data and the time value of the candidate target's location just before the time value of the input sensor data; reducing the first time interval by the amount of time required for the candidate target to change from its status history value just before the time value of the sensor data to the target status inferred for the sensor data; determining a second time interval, wherein the second time interval is the difference between the time value of the candidate target's location that is just after the time value of the sensor data and the time value of the sensor data; reducing the second time interval by the amount of time required for the candidate target to change from the target status inferred for the sensor data to its status history value just after the time value of the sensor data; determining a first speed by dividing the first obstacle avoidance distance by the first reduced time interval; determining a second speed by dividing the second obstacle avoidance distance by the second reduced time interval; and selecting the maximum of the first speed and the second speed as the minimal required speed. 4. The computer readable storage medium of claim 3, wherein the first time interval and the second time interval are shortened using minimum status interval duration statistics and minimum status transition time statistics that are determined from the comparison of a target status history value of the known location of the candidate target at a time value just before a time value of the sensor data and an inferred target status value of the sensor data and a target status history value of the known location of the candidate target at a time value just after a time value of the sensor data and the inferred target status value of the sensor data. 5. The computer readable storage medium of claim 4, wherein the inferred target status value is inferred from the input sensor data using a target class of the candidate target, a geographic region wherein the candidate target is located, and statistical analysis of status histories of candidate targets in the same target class as the candidate target. 6. The computer readable storage medium of claim 1, wherein the step of assigning a statistical weight to each candidate target includes a heuristic comparison of the calculated minimum required speed for each candidate target that is subject to both obstacle avoidance and status transition time constraints with maximum speed capabilities of each candidate target. 7. The computer readable storage medium of claim 1, wherein the calculation of a data association probability for each candidate target to the input sensor data is based upon a ratio of weight determined for the particular candidate target to the combined weight of the initial set of candidate targets. 8. The computer readable storage medium of claim 1, wherein the step of generating a final set of candidate targets includes selecting one or more candidate targets from the set of initial candidate targets wherein the data association probability for each selected candidate target exceeds a threshold value calculated from statistical cluster analysis of the calculated data association probabilities for all of the initial candidate targets. 9. The computer readable storage medium of claim 1, wherein a weight wj for a jth candidate target is based upon a minimal required speed sj=max[s1,s2], where s1 is the speed-of-advance for the candidate target to travel from a point P1 to a point P that is subject to both obstacle avoidance and status transition time constraints, where point P1 is the “just before target history location” for the jth candidate target and where point P is a target location provided by the sensor data, and s2 is the speed-of-advance for the candidate target to travel from the point P to a point P2 that is subject to both obstacle avoidance and status transition time constraints, where point P2 is the “just after target location” for the jth candidate target. 10. The computer readable storage medium of claim 9, wherein if sj≦smax-sustainable, where smax-sustainable is a maximum speed of the jth candidate target that can be sustained over an indefinite period of time, then wj=w1. 11. The computer readable storage medium of claim 9, wherein if smax-sustainable<sj<smax-peak, then w j = w 1 [ S max - peak - s j S max - peak - S max - sustainable ] + w 2 [ s j - S max - sustainable S max - peak - S max - sustainable ] exp ( - Δ t j T j ) . ( 11 ) where w1>w2, where Tj is the peak speed endurance time of the jth candidate target, where smax-peak is the maximum peak speed of the jth candidate target, and where Δtj corresponds to the target status transition adjusted time intervals used to calculate the maximum speed for the jth candidate target. 12. The computer readable storage medium of claim 11, wherein time difference Δt1 is calculated according to Δt1=max[1,T−T1−c(S1,S)], where T represents the time provided by the sensor data, T1 is the time of the “just before location” for the jth initial candidate target, and the status transition time function c(S1,S) is the minimum time required for the jth initial candidate target to transition from its historical target status S1 to the target status S inferred from the sensor data, and where Δt2 is calculated by Δt2=max[1,T2−T−c(S,S2)], where T represents the time provided by the sensor data, T2 is the time corresponding to “just after location” of the jth initial candidate target, and c(S,S2) is the minimum time required for the initial candidate target to transition from the target status S inferred from the sensor data to the historical target status S2 of the jth initial candidate at time T2. 13. The computer readable storage medium of claim 12, wherein the status transition time function is both target-class-dependent and non-symmetric. 14. The computer readable storage medium of claim 9, wherein if sj>smax-peak, then w j = W 2 exp ( - βΔ t j T j ) where β is a target-class dependent tuning parameter and wj=w3 if there is no status and location history available for the jth candidate target, wherein the jth candidate target is a new uniquely identified target with w1>w2>w3. 15. The computer readable storage medium of claim 9, wherein the speed-of-advance s1 is estimated as s 1 = d 1 min Δ t 1 and the speed-of-advance s2 is estimated as s 2 = d 2 min Δ t 2 , where d1min is the minimum obstacle-avoidance distance between P1 and P, and d2min is the minimum obstacle-avoidance distance between P and P2, and where Δt1=max[1,T−T1−c(S1,S)] and Δt2=max[1,T2−T−c(S,S2)]. 16. The computer readable storage medium of claim 15, wherein d1min is calculated by d1min=max[0,doa(P1,P)−u(P1)−u(P)], where doa(P1,P) is the minimum target-class dependent obstacle avoidance distance from point P1 to the point P, u(P1) is determined from the uncertainty of the target location point P1, and u(P) is determined from the uncertainty of the target location point P, and d2min is calculated by d2min=max[0,doa(P,P2)−u(P2)−u(P)], where doa(P, P2) is the target-class dependent obstacle avoidance path between the point P to the point P2 and u(P2) is determined from the uncertainty of the target location. 17. The computer readable storage medium of claim 16, wherein the target-class dependent obstacle avoidance distance calculation utilizes the Floyd-Warshall algorithm. 18. The computer readable storage medium of claim 1, wherein if the initial set of candidate targets comprises m initial candidate targets, total weight wt is defined as the sum of each individual weight w j , w t = ∑ j = 1 m w j . then a data association probability Pj with which a jth candidate target may be associated with the sensor data is determined by p j = w j w t . 19. The computer readable storage medium of claim 18, wherein a set of final candidate targets comprising n final candidate targets is selected from the initial set of candidate targets comprising m initial candidate targets based upon the calculated data association probabilities Pj exceeding a threshold value calculated by statistical cluster analysis from the data association probabilities determined from all initial candidate targets, where m≦n. 20. A system comprising: a display; a processor operatively connected to the display; and a memory module operatively connected to the processor, the memory module having program instructions stored therein, wherein the program instructions are executable by the processor to perform a method comprising the steps of: classifying input sensor data by comparing one or more measured target attributes with one or more known attributes of a finite set of uniquely identified targets; generating an initial set of one or more candidate targets based upon a comparison of the classified input sensor data with the known attributes of the finite set of uniquely identified targets; calculating a minimum required speed for each candidate target to travel between a geolocation history location and a location provided by the sensor data, the minimum required speed subject to both obstacle avoidance and status transition time constraints; assigning a statistical weight to each candidate target based upon its calculated minimum required speed; calculating a data association probability of each candidate target with the classified input sensor data; and generating a final set of candidate targets by selecting the candidate targets from the initial set of candidate targets having calculated data association probabilities that exceed a predetermined threshold value.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.