Light ID error detection and correction for light receiver position determination

• US 9,218,532 B2
• Filed: 09/28/2012
• Issued: 12/22/2015
• Est. Priority Date: 09/28/2012
• Status: Expired due to Fees

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Roberts, Richard D.
• Primary Examiner(s): Bali, Vikkram
• Application Number:
  US13/629,747
• Publication Number:
  US 20140093126A1
• Time in Patent Office: 1,180 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 18/254 of classification results, e.g. of results related to same input data
• CPC Class Codes: G06V 10/24 Aligning, centring, orientation detection or correction of the image
• CPC Class Codes: G06V 10/809 of classification results, e.g. where the classifiers operate on the same input data
• CPC Class Codes: G06V 20/36 Indoor scenes

Abstract

A light receiver records images of light beams originating from a neighborhood of lights, and demodulates identifiers (IDs) from them at determined image positions. The receiver retrieves a set of neighbor IDs for each demodulated ID and a real-world position of the corresponding light. The receiver cross-references the demodulated IDs against the retrieved sets of neighbor IDs to reveal errors in the demodulated IDs. The receiver corrects the errors to produce correct IDs each indexing a real-world position that is correctly matched to one of the determined light beam positions. The receiver determines a position of the receiver relative to the light transmitter based on the correctly matched real-world and determined light beam positions.

Claims

1. A non-transitory computer readable medium encoded with a computer program, including instructions to cause a processor to:

access a recording of images of spatially-separated light beams originating from a neighborhood of lights, each light beam modulated to indicate an identifier (ID) that identifies, and indexes a real-world position of, its originating light;

determine positions of the recorded light beams in the images;

demodulate the ID from each recorded light beam;

retrieve a set of neighbor IDs for each demodulated ID;

cross-reference the demodulated IDs against the sets of neighbor IDs to reveal errors in the demodulated IDs;

correct the errors to produce correct IDs each indexing a real-world position that is correctly matched to one of the determined light beam positions; and

photogrammetrically determine a position of the receiver based on the correctly matched real-world and the determined light beam positions.

2. The computer readable medium of claim 1, wherein the cross-reference reveals:

valid demodulated IDs and any missing IDs which together comprise the correct IDs; and

invalid demodulated IDs.

3. The computer readable medium of claim 2, wherein the cross-reference reveals:

statistical consistencies between the demodulated IDs and the sets of the neighbor IDs that indicate the valid IDs; and

statistical inconsistencies that indicate the missing IDs and invalid demodulated IDs.

4. The computer readable medium of claim 3, wherein the instructions to cross-reference include instructions to cause the processor to:

generate a histogram having

a first axis to represents a union of the demodulated IDs and the sets of neighbor IDs, and

a second axis to represent how many times each ID on the first axis occurs in the sets of neighbor IDs,

wherein first peaks and second peaks in the number of times each ID occurs indicate missing IDs and valid IDs, respectively.

5. The computer readable medium of claim 2, wherein the instructions to correct include instructions to cause the processor to:

retrieve the missing IDs; and

assign each retrieved missing ID to one of the determined light beam positions, from which one of the invalid IDs was demodulated, that is a best fit for the retrieved missing ID.

6. The computer readable medium of claim 5, wherein the instructions to assign include instructions to cause the processor to:

calculate a set of Euclidean distances for each retrieved missing ID relative to all of the valid IDs based on real-world positions corresponding to the retrieved missing ID and the valid IDs; and

assign each retrieved missing ID to one of the determined light beam positions, from which one of the invalid IDs was demodulated, that minimizes the set of Euclidean distances to give the best fit.

7. The computer readable medium of claim 1, wherein the instructions to cause the processor to retrieve include instructions to cause the processor to retrieve a real-world position, in addition to the set of neighbor IDs, for each demodulated ID.

8. A apparatus, comprising:

processing modules to;

access a recording of images of spatially-separated light beams originating from a neighborhood of lights, each light beam modulated to indicate an identifier (ID) that identifies, and indexes a real-world position of, its originating light;

determine positions of the recorded light beams in the images;

demodulate the ID from each recorded light beam;

retrieve a set of neighbor IDs for each demodulated ID;

cross-reference the demodulated IDs against the sets of neighbor IDs to reveal errors in the demodulated IDs;

correct the errors to produce correct IDs each indexing a real-world position that is correctly matched to one of the determined light beam positions; and

photogrammetrically determine a position of the receiver based on the correctly matched real-world and the determined light beam positions.

9. The apparatus of claim 8, wherein the cross-reference reveals:

valid demodulated IDs and any missing IDs which together comprise the correct IDs; and

invalid demodulated IDs.

10. The apparatus of claim 9, wherein the cross-reference further reveals:

statistical consistencies between the demodulated IDs and the sets of the neighbor IDs that indicate the valid IDs; and

statistical inconsistencies that indicate the missing IDs and invalid demodulated IDs.

11. The apparatus of claim 10, wherein the processing modules to cross-reference are configured to generate a histogram having

a first axis to represents a union of the demodulated IDs and the sets of neighbor IDs, and

a second axis to represent how many times each ID on the first axis occurs in the sets of neighbor IDs,

wherein first peaks and second peaks in the number of times each ID occurs indicate missing IDs and valid IDs, respectively.

12. The apparatus of claim 9, wherein the processing modules to correct are configured to:

retrieve the missing IDs; and

assign each retrieved missing ID to one of the determined light beam positions, from which one of the invalid IDs was demodulated, that is a best fit for the retrieved missing ID.

13. The apparatus of claim 12, wherein the processing modules to assign are configured to:

calculate a set of Euclidean distances for each retrieved missing ID relative to all of the valid IDs based on real-world positions corresponding to the retrieved missing ID and the valid IDs; and

assign each retrieved missing ID to one of the determined light beam positions, from which one of the invalid IDs was demodulated, that minimizes the set of Euclidean distances to give the best fit.

14. The apparatus of claim 8, wherein the processing modules to retrieve further include processing modules to retrieve a real-world position, in addition to the set of neighbor IDs, for each demodulated ID.

15. The apparatus of claim 8, further comprising:

a light imager to record the images;

a communication system to communicate with a network;

a processor and memory to interface between the communication system and a user interface system; and

a housing,

wherein the light imager, the processor and memory, and the communication system, are positioned within the housing.

16. The apparatus of claim 15, wherein:

the communication system includes a wireless communication system; and

the housing includes a mobile hand-held housing to house the communication system, the processor, the user interface system, and a battery.

17. A method, comprising:

in a receiver, recording images of spatially-separated light beams originating from a neighborhood of lights, each light beam modulated to indicate an identifier (ID) that identifies, and indexes a real-world position of, its originating light;

determining positions of the recorded light beams in the images;

demodulating the ID from each recorded light beam;

retrieving a set of neighbor IDs for each demodulated ID;

cross-referencing the demodulated IDs against the sets of neighbor IDs to reveal errors in the demodulated IDs;

correcting the errors to produce correct IDs each indexing a real-world position that is correctly matched to one of the determined light beam positions; and

photogrammetrically determining a position of the receiver based on the correctly matched real-world and the determined light beam positions.

18. The method of claim 17, wherein the cross-referencing reveals:

valid demodulated IDs and any missing IDs which together comprise the correct IDs; and

invalid demodulated IDs.

19. The method of claim 18, wherein the cross-referencing reveals:

statistical consistencies between the demodulated IDs and the sets of the neighbor IDs that indicate the valid IDs; and

statistical inconsistencies that indicate the missing IDs and invalid demodulated IDs.

20. The method of claim 19, wherein the cross-referencing includes:

generating a histogram having

a first axis to represents a union of the demodulated IDs and the sets of neighbor IDs, and

a second axis to represent how many times each ID on the first axis occurs in the sets of neighbor IDs,

wherein first peaks and second peaks in the number of times each ID occurs indicate missing IDs and valid IDs, respectively.

21. The method of claim 18, wherein the correcting includes:

retrieving the missing IDs; and

assigning each retrieved missing ID to one of the determined light beam positions, from which one of the invalid IDs was demodulated, that is a best fit for the retrieved missing ID.

22. The method of claim 21, wherein the assigning includes:

calculating a set of Euclidean distances for each retrieved missing ID relative to all of the valid IDs based on real-world positions corresponding to the retrieved missing ID and the valid IDs; and

assigning each retrieved missing ID to one of the determined image positions image positions, from which one of the invalid IDs was demodulated, that minimizes the set of Euclidean distances to give the best fit.

23. The method of claim 17, wherein the retrieving includes retrieving a real-world position, in addition to the set of neighbor IDs, for each demodulated ID.