Back to results
US 2024/0051136 A1Published application
Motion Planning
AssigneeAdaptive Robotics AS
InventorsØystein Hov Holhjem, Gudbrand Eggen, Torstein Anderssen Myhre
Life of this patent
Estimating the term…
Loading patent…
Loading patent…
AssigneeAdaptive Robotics AS
InventorsØystein Hov Holhjem, Gudbrand Eggen, Torstein Anderssen Myhre
Estimating the term…
A method of performing motion planning for a robot in a workspace discretized into workspace elements includes generating or receiving a first model and determining a first set comprising one or more workspace elements that are at least partially in collision with the first model for each of a plurality of states and the respective transition(s) between those states. A first mapping is generated including information regarding the first set and the respective states and transition(s). The method further includes generating or receiving a second model extending from the first model and determining a second set including one or more further workspace elements, additional to those in the first set, that are at least partially in collision with the second model for each of the plurality of states and transitions between those states. A second mapping including information regarding said second set and the respective states and transition(s) is generated.
A method of performing motion planning for a robot, wherein the robot is arranged to operate between a plurality of states in a workspace, said workspace being represented by a plurality of discretized workspace elements, wherein each of said states is connected to at least one other state via at least one respective transition, wherein the method comprises:
The method as claimed in any preceding claim, further comprising:
The method as claimed in any preceding claim, wherein the step of determining the set of workspace elements at least partially overlapped by each model comprises checking the states and transitions that are allowed by the respective dimension of movement afforded by said model.
The method as claimed in any preceding claim, wherein the step of determining the set of workspace elements at least partially overlapped by each model comprises checking the states and transitions for collisions with a static obstacle and/or for self-collisions.
The method as claimed in any preceding claim, wherein at least two of the models are each a representation of the same physical entity.
The method as claimed in any preceding claim, wherein the first and second memory areas are each within a memory, optionally wherein the memory is local.
The method as claimed in any preceding claim, wherein the method is at least partially carried out on a parallel processing unit, optionally wherein the parallel unit comprises a graphical processing unit, optionally wherein the first and second memory areas are each within a memory of said parallel processing unit.
The method as claimed in any preceding claim, wherein one or more mappings are stored in a remote memory or an off-chip memory.
The method as claimed in any preceding claim, further comprising removing one or more workspace elements from the second set that are in first set.
The method as claimed in any preceding claim, wherein one or more workspace elements that are at least partially in collision with more than one of the models are included in the sets corresponding to each of said models.
The method as claimed in any preceding claim, wherein one or more workspace elements that are at least partially in collision with more than one of the models are included in the set corresponding to only one of said models.
The method as claimed in any preceding claim, comprising generating a base mapping comprising information regarding one or more workspace elements that are at least partially in collision with a static part of the robot.
The method as claimed in any preceding claim, wherein the information stored for one or more of the mappings may comprise a plurality of points of interest, optionally wherein one or more of said points of interest comprises a corner and/or a center point.
The method as claimed in any preceding claim, comprising discretizing the workspace into workspace elements of different shapes and/or sizes.
The method as claimed in any preceding claim, further comprising generating a path for the robot to follow using said mappings.
The method as claimed in any preceding claim, wherein one or more of the models is at least partially axially symmetric, the method comprising storing information regarding the workspace elements partially occupied identically for multiple rotational states in a mapping for that model only once.
The method as claimed in any preceding claim, wherein only workspace element hits for one direction of a transition are stored in the corresponding set.
The method as claimed in any preceding claim, wherein only the workspace element hits for a transition that are not found in any of the states to which said transition connects are stored in the corresponding set.
The method as claimed in any preceding claim, wherein workspace element hits that correspond to combined movements are only stored for the higher links, while the simple transitions for the lower links are reused.
The method as claimed in any preceding claim, wherein the states and transitions for first and second dimensions are arranged as a structured roadmap, optionally wherein the structure roadmap is arranged using PDRM.
A motion planner arranged to perform motion planning for a robot, wherein the robot is arranged to operate between a plurality of states in a workspace, said workspace being represented by a plurality of discretized workspace elements, wherein each of said states is connected to at least one other state via at least one respective transition, wherein the motion planner is arranged to:
A non-transitory computer-readable medium comprising instructions that, when executed by a processor, cause the processor to carry out a method of performing motion planning for a robot, wherein the robot is arranged to operate between a plurality of states in a workspace, said workspace being represented by a plurality of discretized workspace elements, wherein each of said states is connected to at least one other state via at least one respective transition, wherein the method comprises:
A method of performing motion planning for a robot, wherein the robot is arranged to operate between a plurality of states in a workspace, said workspace being represented by a plurality of discretized workspace elements, wherein each of said states is connected to at least one other state via at least one respective transition, wherein the method comprises:
A motion planner arranged to perform motion planning for a robot, wherein the robot is arranged to operate between a plurality of states in a workspace, said workspace being represented by a plurality of discretized workspace elements, wherein each of said states is connected to at least one other state via at least one respective transition, wherein the motion planner is arranged to:
A non-transitory computer-readable medium comprising instructions that, when executed by a processor, cause the processor to carry out a method of performing motion planning for a robot, wherein the robot is arranged to operate between a plurality of states in a workspace, said workspace being represented by a plurality of discretized workspace elements, wherein each of said states is connected to at least one other state via at least one respective transition, wherein the method comprises:
Dependencies are inferred from the claim text.