Visual and RF sensor fusion for multi-agent tracking
Methods and systems for determining a device position include determining a first position estimate using radio-based range information. A second position estimate is determined using visual odometry information. The first position estimate and the second position estimate are fused based on radio environmental conditions and visual environmental conditions to determine a final position estimate. Resources are deployed based on the final position estimate.
1. A computer-implemented method for determining a device position, comprising:
determining a first position estimate using radio-based range information;
determining a second position estimate using visual odometry information;
determining an impact of non-line-of-sight paths on the first position estimate;
fusing the first position estimate and the second position estimate based on radio environmental conditions and visual environmental conditions to determine a final position estimate; and
deploying resources based on the final position estimate.
2. The method of claim 1 , wherein determining the impact of non-line-of-sight paths includes determining a radio signal strength of a plurality of anchor devices.
3. The method of claim 2 , wherein determining the impact of non-line-of-sight paths includes determining that fewer than a threshold number of anchor devices have line-of-sight paths.
4. The method of claim 2 , wherein fusing the first position and the second position estimate includes processing a vector that is a concatenation of the first position estimate and the radio signal strength of the plurality of anchor devices.
5. The method of claim 1 , further comprising determining an impact of environmental conditions on the second position estimate.
6. The method of claim 5 , wherein determining the impact of environmental conditions includes determining keypoint matches between image frames.
7. The method of claim 6 , wherein determining the impact of environmental conditions includes determining that the number of keypoint matches is below a threshold value.
8. The method of claim 6 , wherein fusing the first position and the second position estimate includes processing a vector that is a concatenation of the second position estimate and keypoint matches.
9. The method of claim 1 , further comprising determining the radio-based range information using an ultra-wide band transceiver and determining the visual odometry information using a stereo camera.
10. A system for determining a device position, comprising:
a hardware processor; and
a memory that stores a computer program, which, when executed by the hardware processor, causes the hardware processor to:
determine a first position estimate using radio-based range information;
determine a second position estimate using visual odometry information;
determine an impact of non-line-of-sight paths on the first position estimate;
fuse the first position estimate and the second position estimate based on radio environmental conditions and visual environmental conditions to determine a final position estimate; and
deploy resources based on the final position estimate.
11. The system of claim 10 , wherein the computer program further causes the hardware processor to determine a radio signal strength of a plurality of anchor devices.
12. The system of claim 11 , wherein the computer program further causes the hardware processor to determine that fewer than a threshold number of anchor devices have line-of-sight paths.
13. The system of claim 11 , wherein the computer program further causes the hardware processor to determine process a vector that is a concatenation of the first position estimate and the radio signal strength of the plurality of anchor devices for the fusing of the first position estimate and the second position estimate.
14. The system of claim 10 , wherein the computer program further causes the hardware processor to determine an impact of environmental conditions on the second position estimate.
15. The system of claim 14 , wherein the computer program further causes the hardware processor to determine keypoint matches between image frames.
16. The system of claim 15 , wherein the computer program further causes the hardware processor to determine that the number of keypoint matches is below a threshold value.
17. The system of claim 15 , wherein the computer program further causes the hardware processor to process a vector that is a concatenation of the second position estimate and keypoint matches for the fusing of the first position estimate and the second position estimate.
18. The system of claim 10 , further comprising an ultra-wide band transceiver configured to capture the radio-based range information and a stereo camera configured to capture the visual odometry information.
19. A computer-implemented method for determining a device position, comprising:
determining a first position estimate using radio-based range information;
determining a second position estimate using visual odometry information;
determining an impact of environmental conditions on the second position estimate by determining keypoint matches between image frames;
fusing the first position estimate and the second position estimate based on radio environmental conditions and visual environmental conditions to determine a final position estimate, including processing a vector that is a concatenation of the second position estimate and keypoint matches; and
deploying resources based on the final position estimate.