Method and system for robotic task planning
Robots are deployed for handling different tasks in various field of applications. For the robots to function, task planning is required to be done. During the task planning, goal setting is done, as well as actions to be executed for corresponding to each goal are decided. Traditionally, this is carried out first and then the robots start executing the task plan, thereby failing to capture any change in the environment the robots operate, post the task plan generation. Disclosed herein is a method and system for robotic task planning in which a task plan is generated and is executed. However if the task execution fails due to change in any of the parameters/factors, then the system dynamically invokes an adaptation and re-planning mechanism which either updates the already generated task plan (by capturing the change) or generates a new task plan, which the robot can execute to achieve the goal.
1. A system for robotic task planning, comprising:
one or more communication interfaces ( 103 );
one or more hardware processors ( 102 ); and
one or more memory modules ( 101 ) storing a plurality of instructions,
wherein said plurality of instructions, when executed, cause the one or more hardware processors ( 102 ) to:
perform a design time planning to generate a first task plan corresponding to a task;
execute the first task plan using at least one robot in the robotic environment;
fetch a plurality of observations corresponding to the execution of the first task plan;
compare each of the plurality of observations with the first task plan to detect any deviation from the first task plan; and
invoke a re-planning mechanism if any deviation is detected, wherein the re-planning mechanism comprises:
determining whether the deviation is caused due to state change of any existing object, or due to presence of any new object, or due to failure of at least one of a plurality of actions in the first task plan,
wherein
if the deviation is caused due to state change of at least one object:
updating the first task plan to capture the state change of the at least one object; and
executing the updated first task plan using the at least one robot;
if the deviation is caused due to presence of at least one new object:
triggering one or more perception queries to collect details of the at least one new object;
generating a second task plan capturing the details of the at least one new object; and
executing the second task plan using the at least one robot; and
if at least one of a plurality of actions in the first task plan fails:
selecting one of a low granularity approach or a high granularity approach; and
executing the selected low granularity approach or a high granularity approach to update the first task plan.
2. The system as claimed in claim 1 , wherein the robot generates the first task plan by executing the design time planning, comprising:
collecting information pertaining to a planning domain and a planning problem corresponding to the task, as inputs;
performing a heuristic forward/backward search planning on the collected information to determine a plurality of design time plan steps; and
mapping each of the plurality of design time plan steps with corresponding Application Programing Interface (API).
3. The system as claimed in claim 1 , wherein the robot updates the first task plan to capture the state change of the at least one object by:
updating the information pertaining to the planning domain with information pertaining to state change of the at least one object, to generate an updated planning domain;
performing the heuristic forward/backward search planning on the updated planning domain and the planning problem to generate a plurality of updated design time planning steps; and
mapping each of the plurality of updated design time plan steps with corresponding Application Programing Interface (API).
4. A non-transitory computer readable medium for task planning and execution in a robotic environment, wherein the task planning and execution comprising:
performing a design time planning to generate a first task plan corresponding to a task, via one or more hardware processors;
executing the first task plan using at least one robot in the robotic environment, via the one or more hardware processors;
fetching a plurality of observations corresponding to the execution of the first task plan, via the one or more hardware processors;
comparing each of the plurality of observations with the first task plan to detect any deviation from the first task plan, via the one or more hardware processors; and
invoking a re-planning mechanism if any deviation is detected, via the one or more hardware processors, wherein the re-planning mechanism comprises:
determining whether the deviation is caused due to state change of any existing object or due to presence of any new object, or due to failure of at least one of a plurality of actions in the first task plan, wherein
if the deviation is caused due to state change of at least one object:
updating the first task plan to capture the state change of the at least one object; and
executing the updated first task plan using the at least one robot, via the one or more hardware processors;
if the deviation is caused due to presence of at least one new object:
triggering one or more perception queries to collect details of the at least one new object;
generating a second task plan capturing the details of the at least one new object; and
executing the second task plan using the at least one robot, via the one or more hardware processors; and
if at least one of a plurality of actions in the first task plan fails:
selecting one of a low granularity approach or a high granularity approach; and
executing the selected low granularity approach or a high granularity approach to update the first task plan.
5. The non-transitory computer readable medium as claimed in claim 4 , wherein generating the first task plan by performing a design time planning comprises:
collecting information pertaining to a planning domain and a planning problem associated with the task, as inputs;
performing a heuristic forward/backward search planning on the collected information to determine a plurality of design time plan steps; and
mapping each of the plurality of design time plan steps with corresponding Application Programing Interface (API).
6. The non-transitory computer readable medium as claimed in claim 4 , wherein updating the first task plan to capture the state change of the at least one object comprising:
updating the information pertaining to the planning domain with information pertaining to state change of the at least one object, to generate an updated planning domain;
performing the heuristic forward/backward search planning on the updated planning domain and the planning problem to generate a plurality of updated design time planning steps; and
mapping each of the plurality of updated design time plan steps with corresponding Application Programing Interface (API).