Radar map layer in a crowdsourced HD map
Creating and updating an accurate radar map layer for HD map using crowdsourcing may comprise a vehicle obtaining radar data and filtering the radar data on a frame-by-frame basis. In some embodiments, additional filtering may be made on a batch of frames. The vehicle can then transmit the filtered radar data responsive to a determination that a confidence of a position estimate of the vehicle exceeds a conference threshold level and/or a determination that a reliance of the position estimate of the vehicle on the radar data exceeds a reliance threshold level.
1. A method of obtaining data at a vehicle for a radar layer of a map, the method comprising:
obtaining, with a radar at the vehicle, radar data within a geographical region corresponding to at least a portion of the map, the radar data comprising a plurality of frames;
filtering the radar data to remove data generated by moving objects, wherein filtering the radar data comprises identifying, on a per-frame basis, the data generated by the moving objects using:
a Doppler velocity measured by the radar, and
linear and angular velocities of the vehicle when the radar data was obtained, the linear and angular velocities based on vehicle speed determined from a non-radar source; and
selectively transmitting the filtered radar data from the vehicle to at least one device, wherein the transmitting is responsive to:
(i) a determination that a confidence metric, indicative of a confidence level of a 6-DOF position estimate of the vehicle, exceeds a confidence metric threshold level;
(ii) a determination that a reliance metric, indicative of a reliance level of the 6-DOF position estimate of the vehicle on the radar data, exceeds a reliance metric threshold level; or
a combination thereof.
2. The method of claim 1 , further comprising using a calibration transform to transform the linear and angular velocities of the vehicle to a radar frame.
3. The method of claim 1 , further comprising, prior to obtaining the radar data:
sending a notification from the vehicle to a server, wherein the notification comprises an indication of:
a location of the vehicle, and
a capability of the vehicle for obtaining the radar data; and
subsequent to sending the notification, receiving the radar layer of the map at the vehicle from the server.
4. The method of claim 1 , wherein filtering the radar data further comprises filtering a batch of radar data comprising plurality of frames of the radar data.
5. The method of claim 4 , wherein filtering the batch of radar data is based on a clustering algorithm.
6. The method of claim 4 , wherein filtering the batch of radar data comprises spatially identifying and removing outlier radar detections.
7. The method of claim 4 , wherein filtering the batch of radar data comprises:
performing a matching algorithm to determine a degree of similarity of the batch of radar data with corresponding data from the radar layer; and
including the batch of radar data in the filtered radar data responsive to a determination that the degree of similarity of the batch of radar data with the corresponding data from the radar layer is below a threshold level.
8. The method of claim 1 , wherein filtering the radar data further comprises performing random down sampling of the radar data.
9. The method of claim 1 , wherein the determination that the confidence metric of the 6-DOF position estimate of the vehicle exceeds the confidence metric threshold level is based on a covariance matrix of the 6-DOF position estimate.
10. The method of claim 1 , wherein the determination that the reliance of the 6-DOF position estimate of the vehicle on the radar data exceeds the reliance threshold level comprises determining the vehicle has entered a predetermined geographical area.
11. The method of claim 1 , wherein the at least one device comprises a server, another vehicle, or both.
12. The method of claim 1 , further comprising transmitting meta information associated with the filtered radar data, wherein the meta information comprises a matching metric, a statistic of radar localization success, or a sensor quality, or any combination thereof.
13. The method of claim 1 , wherein the radar data comprises point cloud data, data vectorization, or a combination thereof.
14. The method of claim 1 , further comprising transmitting an indication of the confidence of the 6-DOF position estimate of the vehicle.
15. The method of claim 1 , wherein the map comprises a high definition (HD) map.
16. A radar unit for obtaining data at a vehicle for a radar layer of a map, the radar unit comprising:
a radar;
a memory and
one or more processors communicatively coupled with the radar and the memory, wherein the one or more processors are configured to:
obtain, using the radar, radar data within a geographical region corresponding to at least a portion of the map, the radar data comprising a plurality of frames;
filtering the radar data to remove data generated by moving objects, wherein filtering the radar data comprises identifying, on a per-frame basis, the data generated by the moving objects using:
a Doppler velocity measured by the radar, and
linear and angular velocities of the vehicle when the radar data was obtained, the linear and angular velocities based on vehicle speed determined from a non-radar source; and
selectively transmit the filtered radar data from the vehicle to at least one device, wherein the transmitting is responsive to:
(i) a determination that a confidence metric, indicative of a confidence level of a 6-DOF position estimate of the vehicle, exceeds a confidence metric threshold level;
(ii) a determination that a reliance metric, indicative of a reliance level of the 6-DOF position estimate of the vehicle on the radar data, exceeds a reliance metric threshold level; or
a combination thereof.
17. The radar unit of claim 16 , wherein, to filter the radar data on a frame-by-frame basis, the one or more processors are further configured to use a calibration transform to transform the linear and angular velocities of the vehicle to a radar frame.
18. The radar unit of claim 16 , wherein the one or more processors are further configured to, prior to obtaining the radar data:
send a notification from the vehicle to a server, wherein the notification comprises an indication of:
a location of the vehicle, and
a capability of the vehicle for obtaining the radar data; and
subsequent to sending the notification, receive the radar layer of the map at the vehicle from the server.
19. The radar unit of claim 16 , wherein, to filter the radar data, the one or more processors are configured to filter a batch of radar data comprising plurality of frames of the radar data.
20. The radar unit of claim 19 , wherein the one or more processors are configured to filter the batch of radar data based on a clustering algorithm.
21. The radar unit of claim 19 , wherein, to filter the batch of radar data, the one or more processors are configured to spatially identify and remove outlier radar detections.
22. The radar unit of claim 19 , wherein, to filter the batch of radar data, the one or more processors are configured to:
perform a matching algorithm to determine a degree of similarity of the batch of radar data with corresponding data from the radar layer; and
include the batch of radar data in the filtered radar data responsive to a determination that the degree of similarity of the batch of radar data with the corresponding data from the radar layer is below a threshold level.
23. The radar unit of claim 16 , wherein, to filter the radar data, the one or more processors are configured to perform random down sampling of the radar data.
24. The radar unit of claim 16 , wherein the one or more processors are configured to determine that the confidence metric of the 6-DOF position estimate of the vehicle exceeds the confidence metric threshold level based on a covariance matrix of the 6-DOF position estimate.
25. The radar unit of claim 16 , wherein, to determine that the reliance of the 6-DOF position estimate of the vehicle on the radar data exceeds the reliance threshold level, the one or more processors are configured to determine the vehicle has entered a predetermined geographical area.
26. The radar unit of claim 16 , wherein, to transmit the filtered radar data to the at least one device, the one or more processors are configured to send the filtered radar data to a server, another vehicle, or both.
27. The radar unit of claim 16 , wherein the one or more processors are further configured to transmit meta information associated with the filtered radar data, wherein the meta information comprises a matching score, a statistic of radar localization success, or a sensor quality, or any combination thereof.
28. The radar unit of claim 16 , wherein, to obtain the radar data, the one or more processors are configured to obtain point cloud data, data vectorization, or a combination thereof.
29. The radar unit of claim 16 , wherein the one or more processors are further configured to transmit an indication of the confidence of the 6-DOF position estimate of the vehicle.
30. An apparatus for obtaining data at a vehicle for a radar layer of a map, the apparatus comprising:
means for obtaining, at the vehicle, radar data within a geographical region corresponding to at least a portion of the map, the radar data comprising a plurality of frames;
means for filtering the radar data to remove data generated by moving objects, wherein filtering the radar data comprises identifying, on a per-frame basis, the data generated by the moving objects using:
a Doppler velocity measured by the radar, and
linear and angular velocities of the vehicle when the radar data was obtained, the linear and angular velocities based on vehicle speed determined from a non-radar source; and
means for selectively transmitting the filtered radar data from the vehicle to at least one device, wherein the transmitting is responsive to:
(i) a determination that a confidence metric, indicative of a confidence level of a 6-DOF position estimate of the vehicle, exceeds a confidence metric threshold level;
(ii) a determination that a reliance metric, indicative of a reliance level of the 6-DOF position estimate of the vehicle on the radar data, exceeds a reliance metric threshold level; or
a combination thereof.
31. The apparatus of claim 30 , wherein the means for filtering the radar data on a frame-by-frame basis comprise means for using a calibration transform to transform the linear and angular velocities of the vehicle to a radar frame.
32. The apparatus of claim 30 , further comprising:
means for sending, prior to obtaining the radar data, a notification from the vehicle to a server, wherein the notification comprises an indication of:
a location of the vehicle, and
a capability of the vehicle for obtaining the radar data; and
means for receiving, subsequent to sending the notification and prior to obtaining the radar data, the radar layer of the map at the vehicle from the server.
33. The apparatus of claim 30 , wherein the means for filtering the radar data further comprises means for filtering a batch of radar data comprising plurality of frames of the radar data.
34. The apparatus of claim 30 , wherein the means for filtering the radar data further comprises means for performing random down sampling of the radar data.
35. The apparatus of claim 30 , further comprising means for determining the vehicle has entered a predetermined geographical area.
36. The apparatus of claim 30 , wherein the means for selectively transmitting the filtered radar data to the at least one device comprises means for sending the filtered radar data to a server, another vehicle, or both.
37. The apparatus of claim 30 , further comprising means for transmitting meta information associated with the filtered radar data, wherein the meta information comprises a matching score, a statistic of radar localization success, or a sensor quality, or any combination thereof.
38. The apparatus of claim 30 , further comprising means for transmitting an indication of the confidence of the 6-DOF position estimate of the vehicle.
39. A non-transitory computer-readable medium storing instructions for obtaining data at a vehicle for a radar layer of an map, the instructions comprising code for:
obtaining, with a radar at the vehicle, radar data within a geographical region corresponding to at least a portion of the map, the radar data comprising a plurality of frames;
filtering the radar data to remove data generated by moving objects, wherein filtering the radar data comprises identifying, on a per-frame basis, the data generated by the moving objects using:
a Doppler velocity measured by the radar, and linear and angular velocities of the vehicle when the radar data was obtained, the linear and angular velocities based on vehicle speed determined from a non-radar source; and
selectively transmitting the filtered radar data from the vehicle to at least one device, wherein the transmitting is responsive to:
(i) a determination that a confidence metric, indicative of a confidence level of a 6-DOF position estimate of the vehicle, exceeds a confidence metric threshold level;
(ii) a determination that a reliance metric, indicative of a reliance level of the 6-DOF position estimate of the vehicle on the radar data, exceeds a reliance metric threshold level; or
a combination thereof.
40. The computer-readable medium of claim 39 , wherein the instructions further comprise code for, prior to obtaining the radar data:
sending a notification from the vehicle to a server, wherein the notification comprises an indication of:
a location of the vehicle, and
a capability of the vehicle for obtaining the radar data; and
subsequent to sending the notification, receiving the radar layer of the map at the vehicle from the server.
41. The computer-readable medium of claim 39 , wherein the code for filtering the radar data comprises code for filtering a batch of radar data comprising plurality of frames of the radar data.