Robot planning for gaps
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for robot planning. One of the methods includes maintaining a library of pre-generated candidate actions that represent respective robot capabilities; receiving an initial task plan for performing a particular task, wherein the initial task plan includes a sequence of two or more actions that comprises a first action followed by a second action; processing the initial task plan to determine that the first effect of the first action does not result in satisfaction of the second precondition of the second action; selecting one or more selected pre-generated candidate actions from the library of pre-generated candidate actions; and modifying the initial task plan to include the one or more selected pre-generated candidate actions as the intermediate actions to generate a final task plan to be executed by the one or more particular robots when performing the particular task.
1 . A method performed by one or more computers, the method comprising:
maintaining a library of pre-generated candidate actions that represent respective capabilities of one or more particular robots;
receiving an initial task plan for performing a particular task, wherein the initial task plan includes a sequence of two or more actions that comprises a first action followed by a second action, and wherein the first action has a first effect that can be achieved following execution of the first action and the second action has a second precondition that must be satisfied prior to execution of the second action;
processing the initial task plan to determine that the first effect of the first action does not result in satisfaction of the second precondition of the second action, and that the initial task plan is missing one or more intermediate actions between the first action and the second action that are required for performing the particular task with the one or more particular robots, wherein processing the initial task plan comprises iterating through the sequence of two or more actions specified by the initial task plan and keeping track of a latest state of a workcell that includes the one or more particular robots, the latest state of the workcell reflecting the first effect of the first action;
selecting, based on the first effect of the first action and the second precondition of the second action, one or more selected pre-generated candidate actions from the library of pre-generated candidate actions;
modifying the initial task plan to include the one or more selected pre-generated candidate actions as the intermediate actions to generate a final task plan to be executed by the one or more particular robots when performing the particular task; and
controlling the one or more particular robots to perform the particular task based on the final task plan.
2 . The method of claim 1 , wherein the library of pre-generated candidate actions comprises hardware-agonistic actions that represent capabilities shared by different robot models, hardware-specific actions that represent particular capabilities of a particular robot model or particular capabilities with a particular tool, or both.
3 . The method of claim 1 , wherein the one or more intermediate actions comprise a third action that has a third precondition which can be satisfied by the first effect of the first action and a third effect which can result in satisfaction of the second precondition of the second action.
4 . The method of claim 1 , wherein selecting the one or more selected pre-generated candidate actions from the library of pre-generated candidate actions comprises selecting the one or more selected pre-generated candidate actions using a logical model that relates different effects and preconditions to the pre-generated candidate actions.
5 . The method of claim 1 , wherein modifying the initial task plan to include the one more selected pre-generated candidate actions as the intermediate actions comprises using backtracking algorithms to modify one or more preceding actions in the sequence of two or more actions based on the one more selected pre-generated candidate actions.
6 . The method of claim 1 , wherein the initial task plan is provided by a different entity than an entity that generated the library of pre-generated candidate actions.
7 . The method of claim 1 , wherein the initial task plan is agonistic to different robot models.
8 . A system comprising one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:
maintaining a library of pre-generated candidate actions that represent respective capabilities of one or more particular robots;
receiving an initial task plan for performing a particular task, wherein the initial task plan includes a sequence of two or more actions that comprises a first action followed by a second action, and wherein the first action has a first effect that can be achieved following execution of the first action and the second action has a second precondition that must be satisfied prior to execution of the second action;
processing the initial task plan to determine that the first effect of the first action does not result in satisfaction of the second precondition of the second action, and that the initial task plan is missing one or more intermediate actions between the first action and the second action that are required for performing the particular task with the one or more particular robots, wherein processing the initial task plan comprises iterating through the sequence of two or more actions specified by the initial task plan and keeping track of a latest state of a workcell that includes the one or more particular robots, the latest state of the workcell reflecting the first effect of the first action;
selecting, based on the first effect of the first action and the second precondition of the second action, one or more selected pre-generated candidate actions from the library of pre-generated candidate actions;
modifying the initial task plan to include the one or more selected pre-generated candidate actions as the intermediate actions to generate a final task plan to be executed by the one or more particular robots when performing the particular task; and
controlling the one or more particular robots to perform the particular task based on the final task plan.
9 . The system of claim 8 , wherein the library of pre-generated candidate actions comprises hardware-agonistic actions that represent capabilities shared by different robot models, hardware-specific actions that represent particular capabilities of a particular robot model or particular capabilities with a particular tool, or both.
10 . The system of claim 8 , wherein the one or more intermediate actions comprise a third action that has a third precondition which can be satisfied by the first effect of the first action and a third effect which can result in satisfaction of the second precondition of the second action.
11 . The system of claim 8 , wherein selecting the one or more selected pre-generated candidate actions from the library of pre-generated candidate actions comprises selecting the one or more selected pre-generated candidate actions using a logical model that relates different effects and preconditions to the pre-generated candidate actions.
12 . The system of claim 8 , wherein modifying the initial task plan to include the one more selected pre-generated candidate actions as the intermediate actions comprises using backtracking algorithms to modify one or more preceding actions in the sequence of two or more actions based on the one more selected pre-generated candidate actions.
13 . The system of claim 8 , wherein the initial task plan is provided by a different entity than an entity that generated the library of pre-generated candidate actions.
14 . The system of claim 8 , wherein the initial task plan is agonistic to different robot models.
15 . A computer storage medium encoded with a computer program, the program comprising instructions that are operable, when executed by data processing apparatus, to cause the data processing apparatus to perform operations comprising:
maintaining a library of pre-generated candidate actions that represent respective capabilities of one or more particular robots;
receiving an initial task plan for performing a particular task, wherein the initial task plan includes a sequence of two or more actions that comprises a first action followed by a second action, and wherein the first action has a first effect that can be achieved following execution of the first action and the second action has a second precondition that must be satisfied prior to execution of the second action; processing the initial task plan to determine that the first effect of the first action does not result in satisfaction of the second precondition of the second action, and that the initial task plan is missing one or more intermediate actions between the first action and the second action that are required for performing the particular task with the one or more particular robots, wherein processing the initial task plan comprises iterating through the sequence of two or more actions specified by the initial task plan and keeping track of a latest state of a workcell that includes the one or more particular robots, the latest state of the workcell reflecting the first effect of the first action;
selecting, based on the first effect of the first action and the second precondition of the second action, one or more selected pre-generated candidate actions from the library of pre-generated candidate actions;
modifying the initial task plan to include the one or more selected pre-generated candidate actions as the intermediate actions to generate a final task plan to be executed by the one or more particular robots when performing the particular task; and
controlling the one or more particular robots to perform the particular task based on the final task plan.
16 . The computer storage medium of claim 15 , wherein the library of pre-generated candidate actions comprises hardware-agonistic actions that represent capabilities shared by different robot models, hardware-specific actions that represent particular capabilities of a particular robot model or particular capabilities with a particular tool, or both.
17 . The computer storage medium of claim 15 , wherein the one or more intermediate actions comprise a third action that has a third precondition which can be satisfied by the first effect of the first action and a third effect which can result in satisfaction of the second precondition of the second action.
18 . The computer storage medium of claim 15 , wherein selecting the one or more selected pre-generated candidate actions from the library of pre-generated candidate actions comprises selecting the one or more selected pre-generated candidate actions using a logical model that relates different effects and preconditions to the pre-generated candidate actions.
19 . The computer storage medium of claim 15 , wherein modifying the initial task plan to include the one more selected pre-generated candidate actions as the intermediate actions comprises using backtracking algorithms to modify one or more preceding actions in the sequence of two or more actions based on the one more selected pre-generated candidate actions.
20 . The computer storage medium of claim 15 , wherein the initial task plan is provided by a different entity than an entity that generated the library of pre-generated candidate actions.