Systems and methods for utilizing semantic information for navigation of a robotic device

1. A method of improving navigation, the method comprising:

initiating, by a processor in a robotic device, a start of a predetermined time period associated with semantic information extraction;

detecting, during the predetermined period, an adverse event that indicates sensor unreliability of the one or more sensors;

in response to detecting the adverse event that indicates sensor unreliability of the one or more sensors:

identifying a current time slot of the predetermined time period;

identifying a current estimated position and orientation of the robotic device; and

recording updates to semantic information stored for the one or more sensors based on the current time slot and the current estimated position and orientation of the robotic device; and

identifying a pattern of adverse events for a position and orientation of robotic device based on the recorded semantic information after the end of the predetermined time period.

2. The method of claim 1 , further comprising:

in response to determining that the predetermined time period has ended:

identifying, for each sensor, any spatial region of low performance within a local environment of the robotic device based on stored semantic information; and

adjusting a weight factor applied to measurements from the sensor when the robotic device is at an identified spatial region of low performance.

3. The method of claim 2 , further comprising, in response to determining that the predetermined time period has ended:

identifying, for each sensor, any times of low performance based on corresponding semantic information; and

adjusting the weight factor applied to measurements from each sensor during identified times of low performance.

4. The method of claim 2 , wherein identifying, for each sensor, any spatial region of low performance comprises:

determining whether the stored semantic information for the sensor indicates that a number of adverse event occurrences for a specific pose of the robotic device satisfies or exceeds a first event threshold.

5. The method of claim 4 , wherein identifying, for each sensor, any time of low performance comprises determining whether the stored semantic information indicates that a number of adverse event occurrences during a specific timeslot satisfies or exceeds a second event threshold.

6. The method of claim 1 , further comprising for each sensor:

obtaining processed measurements output by pre-processor operations for the sensor;

obtaining the current estimated position and orientation of the robotic device; and

determining whether a difference between a pose indicated by the processed measurements and the current estimated position and orientation satisfies or exceeds a sensor-specific threshold.

7. The method of claim 1 , wherein detecting the adverse event that indicates sensor unreliability of the one or more sensors further comprises determining that the adverse event is detected in response to receiving an outlier data measurement.

8. The method of claim 1 , wherein the predetermined time period associated with semantic information extraction is configured on a time server, wherein the time server maintains settings for timeslots of the predetermined time period.

9. The method of claim 1 , wherein identifying the current estimated position and orientation of the robotic device comprises obtaining output from localization operations on the robotic device, wherein the localization operations are performed as part of a simultaneous localization and mapping (SLAM) process.

10. The method of claim 1 , wherein the robotic device is a ground vehicle, and wherein the one or more sensors of the robotic device comprises at least one image sensor, at least one inertial measurement unit (IMU) sensor, and at least one motion feedback sensor.

11. The method of claim 1 , wherein the semantic information stored for the one or more sensors is represented in a data structure that provides a number of adverse event occurrences for each pose of the robotic device during each timeslot of the predetermined time period.

12. A robotic device, comprising:

one or more sensors; and

a processor coupled to the one or more sensors and configured with processor-executable instructions to:

initiate a start of a predetermined time period associated with semantic information extraction;

detect, during the predetermined period, an adverse event that indicates sensor unreliability of the one or more sensors;

in response to detecting the adverse event that indicates sensor unreliability of the one or more sensors:

identify a current time slot of the predetermined time period;

identify a current estimated position and orientation of the robotic device; and

record updates to semantic information stored for the one or more sensors based on the current time slot and the current estimated position and orientation of the robotic device; and

identify a pattern of adverse events for a position and orientation of robotic device based on the recorded semantic information after the end of the predetermined time period.

13. The robotic device of claim 12 , wherein the processor is further configured with processor-executable instructions to:

in response to determining that the predetermined time period has ended:

identify, for each sensor, any spatial region of low performance within a local environment of the robotic device based on stored semantic information; and

adjust a weight factor applied to measurements from the sensor when the robotic device is at an identified spatial region of low performance.

14. The robotic device of claim 13 , wherein the processor is further configured with processor-executable instructions to:

identify, for each sensor, any times of low performance based on corresponding semantic information in response to determining that the predetermined time period has ended; and

adjust the weight factor applied to measurements from each sensor during identified times of low performance.

15. The robotic device of claim 13 , wherein the processor is further configured with processor-executable instructions to identify, for each sensor, any spatial region of low performance by:

determining whether the stored semantic information for the sensor indicates that a number of adverse event occurrences for a specific pose of the robotic device satisfies or exceeds a first event threshold.

16. The robotic device of claim 15 , wherein the processor is further configured with processor-executable instructions to identify, for each sensor, any time of low performance by determining whether the stored semantic information indicates that a number of adverse event occurrences during a specific timeslot satisfies or exceeds a second event threshold.

17. The robotic device of claim 12 , wherein the processor is further configured with processor-executable instructions to:

obtain processed measurements output by pre-processor operations for each sensor;

obtain the current estimated position and orientation of the robotic device; and

determine whether a difference between a pose indicated by the processed measurements and the current estimated position and orientation satisfies or exceeds a sensor-specific threshold for each sensor.

18. The robotic device of claim 12 , wherein the processor is further configured with processor-executable instructions to detect that the adverse event that indicates sensor unreliability of the one or more sensors in response to receiving an outlier data measurement.

19. The robotic device of claim 12 , wherein the predetermined time period associated with semantic information extraction is configured on a time server, wherein the time server maintains settings for timeslots of the predetermined time period.

20. The robotic device of claim 12 , wherein the processor is further configured with processor-executable instructions to identify the current estimated position and orientation of the robotic device by obtaining output from localization operations on the robotic device, wherein the localization operations are performed as part of a simultaneous localization and mapping (SLAM) process.

21. The robotic device of claim 12 , wherein the robotic device is a ground vehicle, and wherein the one or more sensors comprise at least one image sensor, at least one inertial measurement unit (IMU) sensor, and at least one motion feedback sensor.

22. The robotic device of claim 12 , wherein the semantic information stored for the one or more sensors is represented in a data structure that provides a number of adverse event occurrences for each pose of the robotic device during each timeslot of the predetermined time period.

23. The robotic device of claim 12 , wherein the robotic device is configured to operate within a building.

24. A robotic device, comprising:

means for initiating a start of a predetermined time period associated with semantic information extraction;

means for detecting, during the predetermined period, an adverse event that indicates sensor unreliability of the one or more sensors;

means for identifying a current time slot of the predetermined time period in response to detecting the adverse event that indicates sensor unreliability of the one or more sensors;

means for identifying a current estimated position and orientation of the robotic device in response to detecting the adverse event that indicates sensor unreliability of the one or more sensors;

means for recording updates to semantic information stored for the one or more sensors based on the current time slot and the current estimated position and orientation of the robotic device in response to detecting the adverse event that indicates sensor unreliability of the one or more sensors; and

means for identifying a pattern of adverse events for a position and orientation of robotic device based on the recorded semantic information after the end of the predetermined time period.

25. A processing device for use in a robotic device, the processing device including a processor configured to:

initiate a start of a predetermined time period associated with semantic information extraction;

detect, during the predetermined period, an adverse event that indicates sensor unreliability of the one or more sensors;

in response to detecting the adverse event that indicates sensor unreliability of the one or more sensors the processor is further configured to:

identify a current time slot of the predetermined time period;

identify a current estimated position and orientation of the robotic device; and

record updates to semantic information stored for the one or more sensors based on the current time slot and the current estimated position and orientation of the robotic device; and

identify a pattern of adverse events for a position and orientation of robotic device based on the recorded semantic information after the end of the predetermined time period.

26. The processing device of claim 25 , wherein the processor is further configured to:

in response to determining that the predetermined time period has ended:

identify, for each sensor, any spatial region of low performance within a local environment of the robotic device based on stored semantic information; and

adjust a weight factor applied to measurements from the sensor when the robotic device is at an identified spatial region of low performance.

27. The processing device of claim 26 , wherein the processor is further configured to:

identify, for each sensor, any times of low performance based on corresponding semantic information in response to determining that the predetermined time period has ended; and

adjust the weight factor applied to measurements from each sensor during identified times of low performance.

28. The processing device of claim 26 , wherein the processor is further configured to identify, for each sensor, any spatial region of low performance by:

determining whether the stored semantic information for the sensor indicates that a number of adverse event occurrences for a specific pose of the robotic device satisfies or exceeds a first event threshold.

29. The processing device of claim 28 , wherein the processor is further configured to identify, for each sensor, any time of low performance by determining whether the stored semantic information indicates that a number of adverse event occurrences during a specific timeslot satisfies or exceeds a second event threshold.

30. The processing device of claim 25 , wherein the processor is further configured to:

obtain processed measurements output by pre-processor operations for each sensor;

obtain the current estimated position and orientation of the robotic device; and

determine whether a difference between a pose indicated by the processed measurements and the current estimated position and orientation satisfies or exceeds a sensor-specific threshold for each sensor.

31. The processing device of claim 25 , wherein the processor is further configured to detect the adverse event that indicates sensor unreliability of the one or more sensors in response to receiving an outlier data measurement.

32. The processing device of claim 25 , wherein the predetermined time period associated with semantic information extraction is configured on a time server, wherein the time server maintains settings for timeslots of the predetermined time period.

33. The processing device of claim 25 , wherein the processor is further configured to identify the current estimated position and orientation of the robotic device by obtaining output from localization operations on the robotic device, wherein the localization operations are performed as part of a simultaneous localization and mapping (SLAM) process.

34. The processing device of claim 25 , wherein the robotic device is a ground vehicle, and wherein the one or more sensors comprise at least one image sensor, at least one inertial measurement unit (IMU) sensor, and at least one motion feedback sensor.

35. The processing device of claim 25 , wherein the semantic information stored for the one or more sensors is represented in a data structure that provides a number of adverse event occurrences for each pose of the robotic device during each timeslot of the predetermined time period.