COOPERATIVE MAPPING FOR AUTONOMOUS VEHICLES, ROBOTS OR MULTI-AGENT SYSTEMS

• US 20200135014A1
• Filed: 12/20/2019
• Published: 04/30/2020
• Est. Priority Date: 12/20/2019
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: De La Guardia Gonzalez, Rafael, Aldana Lopez, Rodrigo, Campos Macias, Leobardo, Gomez Gutierrez, David, Guzman Leguel, Anthony Kyung, Parra Vilchis, Jose Ignacio
• Granted Patent: US 11,004,332 B2
• Time in Patent Office: Days
• CPC Class Codes: G01C 21/3811 Point data, e.g. Point of Interest [POI]
• CPC Class Codes: G01C 21/3841 Data obtained from two or more sources, e.g. probe vehicles
• CPC Class Codes: G05D 1/0088 characterized by the autonomous decision making process, e.g. artificial intelligence, predefined behaviours using knowledge based models G06N5/00
• CPC Class Codes: G08G 1/0112 from the vehicle, e.g. floating car data [FCD]
• CPC Class Codes: G08G 1/0145 for active traffic flow control
• CPC Class Codes: G08G 1/161 Decentralised systems, e.g. inter-vehicle communication
• CPC Class Codes: G08G 1/163 involving continuous checking

Abstract

Techniques are disclosed to facilitate cooperative mapping for safe and efficient trajectory planning and collision avoidance by allowing nearby agents to share contextual information. The described techniques also function to extend the mapping range of a single agent by leveraging observations made by multiple agents. Furthermore, the techniques as described herein function to reduce uncertainty in trajectory planning by allowing agents to “see” behind occlusions, thus taking advantage of observations made by neighboring agents from different points of view. An efficient hardware implementation of the system is also presented that leverages the methodologies as discussed herein.

Claims

1. A system on a chip (SoC), comprising:

a control logic subsystem associated with a first agent, the control logic subsystem configured to receive (i) a generated initial point cloud list associated with an online map that is shared with a second agent operating within the same environment as the first agent, the points in the initial point cloud list being mapped based upon sensor data provided by the first agent, and (ii) one or more messages transmitted by the second agent; and

a contextual map subsystem associated with the first agent, the contextual map subsystem being configured to (i) process the one or more messages transmitted by the second agent to identify voxel changes within the generated initial point cloud list identified by the second agent, and (ii) update the points in the initial point cloud list in accordance with the identified voxel changes as indicated in the one or more messages to generate an updated online map.

2. The SoC of claim 1, wherein the identified voxel changes include an object class and at least one attribute of objects contained within each of the identified voxels of the online map.

3. The SoC of claim 1, wherein the identified voxel changes include one or more objects contained within one or more voxels of the online map that are identified by the second agent and unable to be identified via sensors of the first agent.

4. The SoC of claim 1, wherein the one or more messages transmitted by the second agent contain an object of interest (OOI) that indicates the absence or the presence of an obstacle of a certain type associated with the identified voxel changes.

5. The SoC of claim 1, wherein the contextual map subsystem is further configured to select a parent voxel from the initial point cloud list associated with the online map, and to generate an update message for transmission to the second agent by adding obstacles associated with the parent voxel and children of the parent voxel to a voxel list.

6. The SoC of claim 5, wherein the voxel list identifies changes to voxels included in the points in the initial point cloud list associated with the online map that are detected by sensors of the first agent.

7. The SoC of claim 6, further comprising:

a communications subsystem configured to transmit the update message including the obstacles associated with the parent voxel and children of the parent voxel in the voxel list, and

wherein the update message enables the second agent, upon receiving the update message, to update the points in the initial point cloud list in accordance with the identified changes to the voxels in the voxel list to generate a further updated online map.

8. The SoC of claim 1, wherein the first agent is an autonomous vehicle.

9. A non-transitory computer-readable media associated with a first agent having instructions stored thereon that, when executed by one or more processors, cause the first agent to:

map points in an initial point cloud list associated with an online map that is shared with a second agent operating within the same environment as the first agent, the points in the initial point cloud list being mapped based upon sensor data provided by the first agent;

receive one or more messages from the second agent identifying voxel changes included in the online map that was mapped by the first agent; and

update the points in the initial point cloud list in accordance with the identified voxel changes as indicated in the one or more messages to generate an updated online map.

10. The non-transitory computer-readable media of claim 9, wherein the computer-readable media further includes instructions that, when executed by one or more processors, cause the first agent to:

add each of the one or more received messages to a message list; and

update the points in the initial point cloud list in accordance with the identified voxel changes as indicated in each of the one or more messages only when the message list is empty, and otherwise repeatedly mapping obstacles identified in each of the one or more messages contained in the message list.

11. The non-transitory computer-readable media of claim 9, wherein the identified voxel changes include an object class and at least one attribute of objects contained within each of the identified voxels of the online map, and

wherein the one or more messages received from the second agent contain an object of interest (OOI) that indicates the absence or the presence of an obstacle of a certain type associated with the identified voxel changes.

12. The non-transitory computer-readable media of claim 9, wherein the computer-readable media further includes instructions that, when executed by one or more processors, cause the first agent to:

initiate the generation of an update message by selecting a parent voxel from the initial point cloud list associated with the online map; and

generate the update message for transmission to the second agent by adding obstacles associated with the parent voxel and children of the parent voxel to a voxel list,

wherein the voxel list identifies changes to voxels included in the points in the initial point cloud list associated with the online map that are detected by sensors of the first agent.

13. The non-transitory computer-readable media of claim 12, wherein the computer-readable media further includes instructions that, when executed by one or more processors, cause the first agent to:

transmit the update message including the obstacles associated with the parent voxel and children of the parent voxel in the voxel list,

wherein the update message enables the second agent, upon receiving the transmitted update message, to update the points in the initial point cloud list in accordance with the identified changes to the voxels in the voxel list to generate a further updated online map.

14. The non-transitory computer-readable media of claim 13, wherein the computer-readable media further includes instructions that, when executed by one or more processors, cause the first agent to initiate the generation of the update message concurrently with receiving the one or more messages from the second agent.

15. A non-transitory computer-readable media associated with a first agent having instructions stored thereon that, when executed by one or more processors, cause the first agent to:

select a parent voxel from an initial point cloud list associated with an online map that is shared with a second agent operating within the same environment as the first agent, the points in the initial point cloud list being mapped based upon sensor data provided by the first agent;

generate a message for transmission by adding, by the first agent, obstacles associated with the parent voxel and children of the parent voxel to a voxel list that identifies changes to voxels in the online map; and

transmit the message including the obstacles associated with the parent voxel and children of the parent voxel in the voxel list,

wherein the message enables the second agent, upon receiving the message, to update the points in the initial point cloud list in accordance with the identified changes to the voxels in the voxel list to generate an updated online map.

16. The non-transitory computer-readable media of claim 15, wherein the computer-readable media further includes instructions that, when executed by one or more processors, cause the first agent to transmit the message only when the voxel list is empty as a result of no additional changes being identified to the voxels included in the points in the initial point cloud list, and otherwise to repeatedly add the obstacles associated with the parent voxel and children of the parent voxel to the voxel list.

17. The non-transitory computer-readable media of claim 15, wherein the computer-readable media further includes instructions that, when executed by one or more processors, cause the first agent to receive one or more messages from the second agent identifying voxel changes included in the online map that was mapped by the first agent, and to update the points in the initial point cloud list in accordance with the identified voxel changes as indicated in the one or more messages to generate an updated online map.

18. The non-transitory computer-readable media of claim 17, wherein the computer-readable media further includes instructions that, when executed by one or more processors, cause the first agent to add each of the one or more received messages from the second agent to a message list, to update the points in the initial point cloud list in accordance with the identified voxel changes as indicated in each of the one or more messages only when the message list is empty, and to otherwise repeatedly map obstacles identified in each of the one or more messages contained in the message list.

19. The non-transitory computer-readable media of claim 18, wherein the computer-readable media further includes instructions that, when executed by one or more processors, cause the first agent to generate the message concurrently with receiving the one or more messages from the second agent.

20. The non-transitory computer-readable media of claim 15, wherein the identified voxel changes include an object class and at least one attribute of objects contained within each of the identified voxels of the online map, and

wherein the message transmitted to the second agent contains an object of interest (001) that indicates the absence or the presence of an obstacle of a certain type associated with the identified voxel changes.