Autonomous mobile robots for coverage path planning
The disclosure generally relates to a method and a system for heterogeneous autonomous mobile robots for coverage path planning. The method may include receiving sensor data from one or more sensor devices. The sensor data includes information corresponding to one or more robots in a predefined region. The method may further include generating a map for the one or more robots based on the received sensor data. The map includes a probable occupancy of each of the plurality of cells by the one or more robots in the predefined region. The method further includes determining a set of poses of the one or more robots based on the generated map and an optimal set of poses from the set of poses based on the visibility matrix. The method may further include generating a coverage path plan for each of the one or more robots based on the determined optimal set of poses.
1. A processor-implemented method comprising:
receiving, by a processor, sensor data from sensor devices, the sensor data comprising information corresponding to robots in a predefined region, wherein the predefined region is divided into a plurality of cells, wherein the robots are configured to explore the predefined region simultaneously, and broadcast the information dynamically among the robots, based on regions explored as each robot progresses a predefined path;
generating, by the processor, a collision map for the robots based on the received sensor data, the collision map comprising probability of each cell being occupied by each of the robots and visibility of each cell from the plurality of cells accounting for obstruction, wherein the collision map determines feasibility of each robot physically being in a cell among the plurality of cells and wherein the collision map is shared amongst the robots to increase an observable environment for the robots;
identifying, by the processor, a set of goal poses of the robots based on the generated collision map, wherein the set of goal poses and the plurality of cells forms a visibility matrix, the visibility matrix comprising an index for quality of sensor observation;
identifying, by the processor:
an optimal set of goal poses from the set of goal poses based on the visibility matrix, estimating an overall coverage value of each cell in the visibility matrix, and
an optimal order of goal poses visited by the robots from the determined optimal set of goal poses;
generating, by the processor, a coverage path plan for each of the robots based on the identified optimal set of goal poses and the optimal order of poses, wherein generating the coverage path plan comprises assigning a sensor visibility cost to each of the cells corresponding to the optimal set of goal poses;
initiating, by the processor, a replanning for the generated coverage path, wherein the replanning comprises selecting a next goal pose from the set of goal poses;
generating, by the processor, a new coverage path for each of the robots based on said next goal pose; and
facilitating, by the processor, traversal of the robots based on the generated new coverage path to reduce traversed distance by the robots.
2. The method of claim 1 , comprising:
updating coverage path based on feedback received from the sensor data, the feedback comprising the quality of each observation captured by the sensor devices.
3. The method of claim 2 , further comprising: updating probability of quality of each observation captured by the sensor devices.
4. The method of claim 2 , wherein updating the coverage path further comprising:
aggregating the sensor data of each of the robots collected over a period; and
dynamically updating the sensor data upon receiving new information corresponding to the robots.
5. The method of claim 1 , wherein the information corresponding to robots comprises sensor field of view of the robots in the predefined region including the plurality of cells.
6. The method of claim 5 , wherein the sensor field of view comprises position of the robots, orientation of the robots, one or more viewpoint of the robots in the predefined region.
7. The method of claim 1 , comprising:
generating a coverage map, the coverage map comprising a region, from the predefined region, covered by the sensor devices.
8. The method of claim 1 , wherein the visibility matrix comprises assigning an index for quality of sensor observation to each cell in a coverage map, wherein the visibility matrix comprises rows representing the plurality of cell and columns representing the set of goal poses.
9. The method of claim 1 , wherein identifying the optimal order of goal poses comprises creating a distance metrics for each pose from the optimal set of goal poses accounting for kinematic constraints and feasibility of each of the robots, and wherein identifying the optimal order of goal poses enables prioritizing poses to be visited by the robots.
10. The method of claim 1 , wherein the sensor visibility cost assigned to each of the cells is indicative of a measure of confidence of the visibility of the cell by each of the robots.
11. A system comprising:
a memory storing instructions;
a processor coupled to the memory, wherein the processor is configured by the instructions to:
receive sensor data from sensor devices, the sensor data comprising information corresponding to robots in a predefined region, wherein the predefined region is divided into a plurality of cells, wherein the robots are configured to explore the predefined region simultaneously, and broadcast the information dynamically among the robots, based on regions explored as each robot progresses a predefined path;
generate a collision map for the robots based on the received sensor data, the collision map comprising probability of each cell being occupied by each of the robots and visibility of each cell from the plurality of cells accounting for obstruction, wherein the collision map determines feasibility of each robot physically being in a cell among the plurality of cells and wherein the collision map is shared amongst the robots to increase an observable environment for the robots;
identify a set of goal poses of the robots based on the generated collision map, wherein the set of goal poses and the plurality of cells forms a visibility matrix, the visibility matrix comprising an index for quality of sensor observation;
identify:
an optimal set of goal poses from the set of goal poses based on the visibility matrix estimating an overall coverage value of each cell in the visibility matrix, and
an optimal order of poses visited by the robots from the determined optimal set of goal poses;
generate a coverage path plan for each of the robots based on the identified optimal set of goal poses and the optimal order of poses, wherein generating the coverage path plan comprises assigning a sensor visibility cost to each of the cells corresponding to the optimal set of goal poses;
initiate a replanning for the generated coverage path, wherein the replanning comprises selecting a next goal pose from the set of goal poses;
generate a new coverage path for each of the robots based on said next goal pose; and
facilitate traversal of the robots based on the generated new coverage path to reduce traversed distance by the robots.
12. The system of claim 11 , further configured to:
updating coverage path based on feedback received from the sensor data, the feedback comprising the quality of each observation captured by the sensor devices.
13. The system of claim 12 , further configured to update probability of quality of each observation captured by the sensor devices.
14. The system of claim 12 , further configured to:
aggregate the sensor data of each of the robots collected over a period; and
dynamically update the sensor data upon receiving new information corresponding to the robots.
15. The system of claim 11 , wherein the information corresponding to robots comprises sensor field of view of the robots in the predefined region including the plurality of cells.
16. The system of claim 15 , wherein the sensor field of view comprises position of the robots, orientation of the robots, one or more viewpoint of the robots in the predefined region.
17. The system of claim 11 , further configured to:
generate a coverage map, the coverage map comprising a region, from the predefined region, covered by the sensor devices.
18. The system of claim 11 , wherein the visibility matrix comprises:
assigning an index for quality of sensor observation to each cell in a coverage map, wherein the visibility matrix comprises rows representing the plurality of cell and columns representing the set of goal poses.
19. The system of claim 11 , wherein to identify the optimal order of goal poses, the processor is further configured to:
create a distance metrics for each pose from the optimal set of goal poses accounting for kinematic constraints and feasibility of each of the robots, and wherein identifying the optimal order of goal poses enables prioritizing poses in which the poses are to be visited by the robots.
20. The system of claim 11 , wherein the sensor visibility cost assigned to each of the cells is indicative of a measure of confidence of the visibility of the cell by each of the robots.
21. One or more non-transitory machine-readable information storage mediums comprising one or more instructions which when executed by one or more hardware processors cause:
receiving sensor data from sensor devices, the sensor data comprising information corresponding to robots in a predefined region, wherein the predefined region is divided into a plurality of cells, wherein the robots are configured to explore the predefined region simultaneously, and broadcast the information dynamically among the robots, based on regions explored as each robot progresses a predefined path;
generating a collision map for the robots based on the received sensor data, the collision map comprising probability of each cell being occupied by each of the robots and visibility of each cell from the plurality of cells accounting for obstruction, wherein the collision map determines feasibility of each robot physically being in a cell among the plurality of cells and wherein the collision map is shared amongst the robots to increase an observable environment for the robots;
identifying a set of goal poses of the robots based on the generated map, wherein the set of goal poses and the plurality of cells forms a visibility matrix, the visibility matrix comprising an index for quality of sensor observation;
identifying:
an optimal set of goal poses from the set of goal poses based on the visibility matrix, estimating an overall coverage value of each cell in the visibility matrix, and
an optimal order of poses visited by the robots from the determined optimal set of goal poses;
generating a coverage path plan for each of the robots based on the identified optimal set of goal poses and the optimal order of poses, wherein generating the coverage path plan comprises assigning a sensor visibility cost to each of the cells corresponding to the optimal set of goal poses;
initiating a replanning for the generated coverage path, wherein the replanning comprises selecting a next goal pose from the set of goal poses;
generating a new coverage path for each of the robots based on said next goal pose; and
facilitating, by the processor, traversal of the robots based on the generated new coverage path to reduce traversed distance by the robots.