Computer-implemented method and system for locating an indoor object

• US 8,325,090 B2
• Filed: 07/23/2009
• Issued: 12/04/2012
• Est. Priority Date: 05/19/2009
• Status: Active Grant

• Current Assignee: National Taiwan University (Government of The Republic of China)
• Original Assignee: National Taiwan University (Government of The Republic of China)
• Inventors: Luo, Ren-Chyuan, Chen, Chien-Chung
• Primary Examiner(s): Issing, Gregory C
• Application Number:
  US12/508,040
• Publication Number:
  US 20100295733A1
• Time in Patent Office: 1,230 Days
• Field of Search: 342/450, 342/451
• US Class Current: 342/451
• CPC Class Codes: G01S 13/825 with exchange of information between interrogator and responder
• CPC Class Codes: G01S 13/86 Combinations of radar systems with non-radar systems, e.g. sonar, direction finder
• CPC Class Codes: G01S 13/878 Combination of several spaced transmitters or receivers of known location for determining the position of a transponder or a reflector G01S13/874 takes precedence
• CPC Class Codes: G01S 5/0264 at least one of the systems being a non-radio wave positioning system

Abstract

A computer-implemented method for locating an indoor object includes: a) receiving an RF coordinate signal and a detecting signal; b) providing information of a located region with reference to the RF coordinate and the detecting signal; c) obtaining a motion sensor coordinate according to the information of the located region; d) providing weights for the RF coordinate and the motion sensor coordinate; e) applying the weights to the RF coordinate and the motion sensor coordinate; and f) combining the weighted RF coordinate and the weighted motion sensor coordinate to generate a fused coordinate corresponding to the position of the indoor object. A system for locating an indoor object is also disclosed.

Claims

1. A computer-implemented method for locating an indoor object based on a radio frequency (RF) position signal from an RF position estimation system, and at least one detecting signal from a motion sensor array upon detection of movement of the indoor object, the RF position signal carrying information of an RF position on a position plane estimating a position of the indoor object, said method comprising:

a) receiving the RF position signal and the detecting signal;

b) providing information of a located region on the position plane with reference to the RF position and the detecting signal;

c) obtaining a motion sensor position on the position plane according to the information of the located region;

d) adopting a covariance intersection (CI) estimator for offline training and generating weights for the RF position and the motion sensor position;

e) applying the weights to the RF position and the motion sensor position, respectively; and

f) combining the weighted RF position and the weighted motion sensor position to generate a fused position on the position plane corresponding to the position of the indoor object;

wherein when only one detecting signal is received in step a), step b) is performed by providing information of a predetermined sensing area of the motion sensor array on the position plane in which the RF position falls, followed by defining the predetermined sensing area as the located region; and

wherein when more than one detecting signals are received in step a), step b) includes the sub-steps of;

b-1) providing information of a plurality of pre-determined sensing areas of the motion sensor array on the position plane, such that at least two of the pre-determined sensing areas overlap each other so as to divide the pre-determined sensing areas into an overlapped region and non-overlapped regions; and

b-2) determining one of the non-overlapped regions and the overlapped region, in which the RF position falls, as the located region.

2. The method of claim 1 , wherein the motion sensor position is a geometric center of the predetermined sensing area.

3. The method of claim 1, wherein the motion sensor position is a geometric center of one of the predetermined sensing areas corresponding to one of the non-overlapped regions when the RF position falls in said one of the non-overlapped regions, and is a geometric center of the overlapped region when the RF position falls in the overlapped region.

4. The method of claim 1, wherein, in step d), a sum of the weights is equal to 1, and in step f), the weighted RF position and the weighted sensor position are summed to generate the fused position.

5. The method of claim 1, wherein step d) includes the sub-steps of:

d-1) providing a plurality of actual positions corresponding to a plurality of training patterns, respectively;

d-2) providing a plurality of RF training positions corresponding to the training patterns, respectively;

d-3) computing an RF position covariance matrix according to the actual positions and the RF training positions corresponding to the training patterns;

d-4) providing a plurality of motion sensor training positions corresponding to the training patterns, respectively;

d-5) computing a motion sensor position covariance matrix according to the actual positions and the motion sensor training positions corresponding to the training patterns; and

d-6) generating the weights with the CI estimator according to the RF position covariance matrix and the motion sensor position covariance matrix.

6. The method of claim 5, wherein sub-step d-6) includes the sub-steps of:

d-6-1) giving a training weight W;

d-6-2) computing a fused position covariance matrix with the CI estimator according to the training weight W, the RF position covariance matrix, and the motion sensor position covariance matrix;

d -6-3) performing d-6-1) and d-6-2) iteratively until a minimum trace value for the fused position covariance matrix is found; and

d-6-4) taking the last given training weight W and (1-W) as the weights when the minimum trace value of the fused position covariance matrix is found.

7. The method of claim 5, wherein sub-step d-6) includes the sub-steps of:

d-6-1) giving a training weight W;

d-6-2) computing a fused position covariance matrix with the CI estimator according to the training weight W, the RF position covariance matrix, and the motion sensor position covariance matrix;

d-6-3) performing d-6-1) and d-6-2) iteratively until a minimum determinant value for the fused coordinate covariance matrix is found; and

d-6-4) taking the last given training weight W and (1-W) as the weights when the minimum determinant value of the fused position covariance matrix is found.

8. A system for locating an indoor object, comprising:

a radio frequency (RF) position estimation system including at least three reference node RF transceivers, a blind node RF transceiver, and a localization module, wherein each of said reference node RF transceivers correspond to a pre-determined reference node coordinate on a position plane and are placed at a pre-determined position, said blind node RF transceiver being carried by an indoor object, said localization module being configured to receive Received Signal Strength Index (RSSI) values related to distances between said at least three reference node RF transceivers and said blind node RF transceiver, and to generate a RF position on the position plane according to the pre-determined reference node positions on the position plane and the RSSI values;

a motion sensor position estimation system including a motion sensor array and a localization module, said motion sensor array including at least one motion sensor configured to output a detecting signal upon detection of movement of the indoor object, said localization module of the motion sensor position estimation system being configured to receive the RF position signal and the detecting signal, to provide information of a located region on the position plane with reference to the RF position and the detecting signal, and to obtain a motion sensor position on the position plane according to the information of the located region; and

a fusion system including a weight generation module and a fused position calculation module, said weight generation module of said fusion system adopting a covariance intersection (CI) estimator for offline training and generating the weights for the RF position and the motion sensor position, said fused position calculation module being configured to receive the RF position and the motion sensor position, to apply the weights to the RF position and the motion sensor position, respectively, and to combine the weighted RF position and the weighted motion sensor position to generate a fused position on the position plane corresponding to the position of the indoor object.

9. The system of claim 8, wherein said motion sensor array of said motion sensor position estimation system includes a motion sensor, said localization module being configured to provide information of a pre-determined sensing area of the motion sensor array on the position plane in which the RF position falls, and to define the pre-determined sensing area.

10. The system of claim 9, wherein the motion sensor position obtained from said localization module according to the information of the predetermined sensing area of the motion sensor array on the position plane in which the RF position falls is a geometric center of the pre-determined sensing area.

11. The system of claim 8, wherein said motion sensor array of said motion sensor position estimation system includes a plurality of motion sensors, said localization module of said motion sensor position estimation system being configured to provide information of a plurality of pre-determined sensing areas of the motion sensor array on the position plane in response to the detecting signals generated from said motion sensors, such that at least two of the pre-determined sensing areas overlap each other so as to divide the pre-determined sensing areas into an overlapped region and non-overlapped regions, said localization module being further configured to determine one of the non-overlapped regions and the overlapped region, in which the RF position falls.

12. The system of claim 11, wherein said localization module of said motion sensor position estimation system is further configured in such a manner that when the RF position falls in one of the non-overlapped regions, the motion sensor position obtained from said localization module according to the information of the located region is a geometric center of one of the predetermined sensing areas corresponding to said one of the non-overlapped regions, and is a geometric center of the overlapped region when the RF position falls in the overlapped region.

13. The system of claim 8, wherein a sum of the weights provided by said weight generation module of said fusion system is equal to 1, and said fused position calculation module is configured to sum the weighted RF position and the weighted motion sensor position to generate the fused position.

14. The system of claim 8, wherein said weight generation module of said fusion system is configured to compute an RF position covariance matrix according to a plurality of actual positions corresponding to a plurality of training patterns, respectively, and a plurality of RF training positions corresponding to the training patterns, respectively, further to compute a motion sensor position covariance matrix according to the actual positions and a plurality of motion sensor training positions corresponding to the training patterns, respectively, and to generate the weights with the CI estimator according to the RF position covariance matrix and the motion sensor position covariance matrix.