Robot motion planning
AssigneeKUKA Hungária Korlátolt Felelösségû Társaság
InventorIstván Komlósi
Life of this patent
Estimating the term…
Abstract
A method for motion planning for at least one robot includes providing a start configuration comprising at least one start position and a destination configuration comprising at least one destination position for the robot, providing a motion of at least one obstacle in the workspace of the robot, the obstacle motion defining a position of the obstacle that varies over time, and determining a motion of the robot from its start configuration to its destination configuration. The robot motion definies a position of the robot over a time period from a start time to a destination time. The robot motion is determined such that at each point in time between the start and destination times a distance between the robot and the obstacle does not fall below a predetermined threshold.
Claims16
A method for motion planning for at least one robot having a plurality of axes that are actuated by respective drives and a microprocessor adapted to control the drives, the method comprising:
- providing a start configuration of a first robot comprising at least one start position;
- providing a destination configuration of the first robot comprising at least one destination position;
- providing information for a motion of at least one obstacle in the workspace of the first robot, the obstacle motion information defining a position of the obstacle varying over time;
- determining with the microprocessor a motion of the first robot from the start configuration to the destination configuration, the robot motion defining a position of the first robot over a time period from a start time to a destination time;
- wherein the robot motion is determined such that at each point in time between the start time and the destination time, a distance between the first robot and the obstacle does not fall below a predetermined threshold;
- wherein determining the motion of the first robot comprises employing at least one of:
- a state space comprising a time dimension and at least one position dimension, or
- a hyperspace comprising state transitions of the state space;
- wherein at least one of the state space or the hyperspace is discretized;
- determining a first motion of the first robot employing at least one of the state space or the hyperspace discretized at a first resolution;
- and
- determining a subsequent motion of the first robot employing a subspace of at least one of the state space or the hyperspace, wherein the subspace is at least one of in the vicinity of the first motion, or is discretized at a subsequent resolution finer than the first resolution.
A system for planning motion of at least one robot, the system comprising:
- means for providing a start configuration of the robot, the start configuration comprising at least one start position of the robot, and for providing at least one destination configuration of the robot that comprises at least one destination position for the robot;
- means for providing information of a motion of at least one obstacle in the workspace of the robot, the obstacle motion defining a position of the obstacle varying over time;
- and
- means for determining a motion of the robot from the start configuration to the destination configuration, the robot motion defining a position of the robot over a time period from a start time to a destination time, such that at each point in time between the start time and the destination time, a distance between the robot and the obstacle does not fall below a predetermined threshold;
- wherein determining the motion of the robot comprises employing at least one of:
- a state space comprising a time dimension and at least one position dimension, or
- a hyperspace comprising state transitions of the state space;
- wherein at least one of the state space or the hyperspace is discretized;
- and
- wherein determining the motion of the robot further comprises:
- determining a first motion of the robot employing at least one of the state space or the hyperspace discretized at a first resolution, and
- determining a subsequent motion of the robot employing a subspace of at least one of the state space or the hyperspace, wherein the subspace is at least one of in the vicinity of the first motion, or is discretized at a subsequent resolution finer than the first resolution.
A computer program product comprising source code stored on a non-transitory computer-readable medium, the source code, when executed on a microprocessor, causing the microprocessor to:
- provide a start configuration of a first robot, the start configuration comprising at least one start position;
- provide a destination configuration of the first robot, the destination configuration comprising at least one destination position;
- provide information for a motion of at least one obstacle in the workspace of the first robot, the obstacle motion information defining a position of the obstacle varying over time;
- and
- determine a motion of the first robot from the start configuration to the destination configuration, the robot motion defining a position of the first robot over a time period from a start time to a destination time;
- wherein the robot motion is determined such that at each point in time between the start time and the destination time, a distance between the first robot and the obstacle does not fall below a predetermined threshold;
- wherein determining the motion of the robot comprises employing at least one of:
- a state space comprising a time dimension and at least one position dimension, or
- a hyperspace comprising state transitions of the state space;
- wherein at least one of the state space or the hyperspace is discretized;
- and
- wherein determining the motion of the robot further comprises:
- determining a first motion of the robot employing at least one of the state space or the hyperspace discretized at a first resolution, and
- determining a subsequent motion of the robot employing a subspace of at least one of the state space or the hyperspace, wherein the subspace is at least one of in the vicinity of the first motion, or is discretized at a subsequent resolution finer than the first resolution.
Dependencies are inferred from the claim text.