Techniques for validating UAV position using visual localization
Systems and methods for validating a position of an unmanned aerial vehicle (UAV) are provided. A method can include receiving map data for a location, the map data including labeled data for a plurality of landmarks in a vicinity of the location. The method can include generating image data for the location, the image data being derived from images of the vicinity generated by the UAV including at least a subset of the plurality of landmarks. The method can include determining a visual position of the UAV using the image data and the map data. The method can include determining a Global Navigation Satellite System (GNSS) position of the UAV. The method can include generating an error signal using the visual position and the GNSS position. The method can also include validating the GNSS position in accordance with the error signal satisfying a transition condition.
1. A computer-implemented method for validating a position of an unmanned aerial vehicle (UAV), the method comprising:
receiving map data for a location, the map data comprising labeled data for a plurality of landmarks in a vicinity of the location;
generating image data for the location, the image data being derived from images of the vicinity generated by the UAV including at least a subset of the plurality of landmarks;
determining a visual position of the UAV using the image data and the map data;
determining a Global Navigation Satellite System (GNSS) position of the UAV;
generating an error signal using the visual position and the GNSS position; and
validating the GNSS position in accordance with the error signal satisfying a transition condition that is used to determine when to localize the UAV based on the visual position or the GNSS position,
wherein generating the error signal comprises generating an error vector using the visual position and the GNSS position, and wherein the transition condition comprises a threshold magnitude of the error vector above which the GNSS position is deemed invalid.
2. The computer-implemented method of claim 1 , wherein generating the error signal comprises generating a parity vector based at least in part on visual position data and GNSS position data, and wherein the transition condition comprises a threshold magnitude of the parity vector above which the GNSS position is invalid.
3. The computer-implemented method of claim 2 , wherein the visual position data comprise multiple feature measurements from multiple features in the vicinity and the GNSS position data comprise multiple GNSS measurements from multiple satellites of the GNSS system, and wherein generating the parity vector comprises defining a parity space using the feature measurements and the GNSS measurements and determining the magnitude of the parity vector in the parity space.
4. The computer implemented method of claim 2 , further comprising determining a magnitude of noise in the parity vector, wherein the threshold magnitude of the parity vector is about equal to the magnitude of noise.
5. The computer implemented method of claim 2 , wherein:
the visual position of the UAV is a first visual position of the UAV;
determining the visual position of the UAV comprises generating the first visual position using a first technique and generating a second visual position using a second technique different from the first technique;
generating the error signal comprises generating the parity vector based at least in part on the first visual position, the second visual position, and the GNSS position; and
the transition condition further comprises a comparative condition for which GNSS data is invalid where the parity vector is more closely oriented toward a first coordinate axis in a parity space corresponding to the GNSS position than toward a second coordinate axis in the parity space corresponding to the first visual position or the second visual position.
6. The computer implemented method of claim 2 , wherein the parity vector is a first parity vector and the magnitude is a first magnitude, the method further comprising:
generating altitude data using two different altitude sensors of the UAV;
determining a GNSS altitude of the UAV;
generating a second parity vector based at least in part on the GNSS altitude and the altitude data; and
isolating a source of sensor error to the altitude data or the GNSS position, based at least in part on a second value of the second parity vector.
7. The computer implemented method of claim 1 , wherein the transition condition comprises a minimum altitude below which the GNSS position is invalid.
8. The computer implemented method of claim 1 , further comprising:
initiating an ascent of the UAV from the location; and
discontinuing visual localization in accordance with the error signal satisfying the transition condition.
9. The computer implemented method of claim 1 , further comprising:
initiating a descent of the UAV toward the location; and
transitioning from GNSS localization to visual localization in accordance with the error signal failing to satisfy the transition condition.
10. The computer implemented method of claim 1 , wherein generating the error signal comprises determining a temporal variation of the GNSS position and wherein the transition condition comprises a threshold for the temporal variation above which the GNSS position is invalid.
11. The computer implemented method of claim 1 , wherein generating the error signal comprises determining a temporal variation of the visual position and wherein the transition condition comprises a threshold for the temporal variation above which the visual position is invalid.
12. The computer implemented method of claim 1 , further comprising:
modifying a mission of the UAV in accordance with GNSS data failing to satisfy the transition condition.
13. At least one machine-accessible storage medium onboard an unmanned aerial vehicle (UAV) storing instructions that, when executed by circuitry of the UAV, will cause the circuitry to perform operations comprising:
receiving map data for a location, the map data comprising labeled data for a plurality of landmarks in a vicinity of the location;
generating image data for the location, the image data being derived from images of the vicinity generated by the UAV including at least a subset of the plurality of landmarks;
determining a visual position of the UAV using the image data and the map data;
determining a GNSS position of the UAV;
generating an error signal using the visual position and the GNSS position; and
validating the GNSS position in accordance with the error signal satisfying a transition condition that is used to determine when to localize the UAV based on the visual position or the GNSS position,
wherein generating the error signal comprises generating a parity vector based at least in part on visual position data and the GNSS position data, wherein the transition condition comprises a threshold magnitude of the parity vector above which the GNSS position is invalid.
14. The at least one machine-accessible storage medium of claim 13 , wherein generating the error signal further comprises generating an error vector using the visual position and the GNSS position, wherein the transition condition comprises a threshold magnitude of the error vector above which the GNSS position is invalid.
15. The at least one machine-accessible storage medium of claim 13 , wherein the visual position data comprise multiple feature measurements from multiple landmarks in the vicinity and the GNSS position data comprise multiple GNSS measurements from multiple satellites of the GNSS system, and wherein generating the parity vector comprises defining a parity space using the feature measurements and the GNSS measurements and determining a magnitude of the parity vector in the parity space.
16. The at least one machine-accessible storage medium of claim 13 , further comprising determining a magnitude of noise in the parity vector, wherein the threshold magnitude of the parity vector is about equal to the magnitude of noise.
17. The at least one machine-accessible storage medium of claim 13 , wherein:
the visual position of the UAV is a first visual position of the UAV;
determining the visual position of the UAV comprises generating the first visual position using a first technique and generating a second visual position using a second technique different from the first technique;
generating the error signal comprises generating the parity vector based at least in part on the first visual position, the second visual position, and the GNSS position; and
the transition condition further comprises a comparative condition for which GNSS data is invalid where the parity vector is more closely oriented toward a first coordinate axis in a parity space corresponding to the GNSS position that toward a second coordinate axis in the parity space corresponding the first visual position or the second visual position.
18. The at least one machine-accessible storage medium of claim 13 , wherein generating the error signal further comprises determining a number of unique GNSS signals included in the GNSS data, wherein the transition condition comprises a minimum number of unique GNSS signals included in the GNSS data, below which the GNSS data is invalid.
19. The at least one machine-accessible storage medium of claim 13 , wherein the transition condition comprises a minimum altitude below which the GNSS position is invalid.
20. The at least one machine-accessible storage medium of claim 13 , wherein generating the error signal comprises determining a temporal variation of the GNSS position and wherein the transition condition comprises a threshold for the temporal variation above which the GNSS position is invalid.