Method and system for task execution in dynamic heterogeneous robotic environment
A method and system for task execution in dynamic heterogeneous robotic environment is disclosed. The method includes extracting data associated with a plurality of data categories and further deriving a plurality of factors from the extracted data. The method further includes determining a plurality of correlations among the plurality of factors based on the deep learning network. The method further includes deriving a plurality of sentiment parameters for a set of factors from the plurality of factors based on the plurality of correlations. The method may further includes simulating execution of the at least one current task by employing a plurality of robots. The method may further includes iteratively re-adjusting at least one of the plurality of sentiment parameters based on reinforcement learning performed on a result of the simulating. The method may further includes executing the at least one current task by employing the plurality of robots.
1. A method for task execution in dynamic heterogeneous robotic environment, the method comprising:
extracting, by a task execution device, data associated with a plurality of data categories in response to identifying at least one current task, wherein the data is extracted from a plurality of data sources, and wherein each of the plurality of data categories is associated with at least one of the plurality of data sources;
deriving, by the task execution device, a plurality of factors from the extracted data based on a deep learning network;
determining, by the task execution device, a plurality of correlations among the plurality of factors based on the deep learning network, wherein the plurality of correlations comprise weighted relationship between two or more factors from the plurality of factors;
deriving, by the task execution device, a plurality of sentiment parameters for a set of factors from the plurality of factors based on the plurality of correlations, wherein the plurality of sentiment parameters are hierarchical;
simulating, by the task execution device, execution of the at least one current task by employing a plurality of robots based on each of the plurality of sentiment parameters applied to at least one associated factor from the set of factors;
iteratively re-adjusting, by the task execution device, at least one of the plurality of sentiment parameters based on reinforcement learning performed on a result of the simulating, wherein the iterative re-adjustment is performed until each of the plurality of sentiment parameters are above a predefined threshold; and
executing, by the task execution device, the at least one current task by employing the plurality of robots, based on each of the plurality of sentiment parameters applied to at least one associated factor from the set of factors in response to the iterative re-adjusting.
2. The method of claim 1 further comprising receiving the at least one current task requiring execution from a task pipeline.
3. The method of claim 2 , wherein the extracting the data comprises:
determining a current state of the task execution device, in response to receiving the at least one current task; and
identifying the data based on the current state of the task execution device and the at least one current task received.
4. The method of claim 1 , wherein the plurality of data sources comprises an environmental factors database and a historical database, wherein the historical database comprises at least one of order information, factory functioning information, external information, robot history information, or task handling information.
5. The method of claim 1 , wherein the plurality of factors comprises at least one of date, day, time, product visibility, product character, or product category, and wherein each of the plurality of factors represents a neuron in the deep learning network.
6. The method of claim 1 , wherein the determining the plurality of correlations among the plurality of factors comprises:
assigning weights to relations between a set of factors from the plurality of factors to generated weighted relationship between the set of factors;
analyzing the assigned weights based on a deep neural network; and
re-adjusting the assigned weights based on a result of the analyzing.
7. The method of claim 1 , wherein the hierarchy associated with a sentiment parameter of the plurality of sentiment parameters corresponds to at least one of roles, task allocation to the plurality of robots, robot functioning, or product categories.
8. The method of claim 1 , wherein the simulating execution of the at least one current task comprises:
determining a plurality of use cases for simulating execution of the at least one current task, wherein each of the plurality of use cases comprises application of each of the plurality of sentiment parameters to at least one associated factor from the set of factors; and
assigning, for each of the plurality of use cases, at least one reward point to each of the plurality of robots, based on simulation of the at least one current task.
9. The method of claim 8 , wherein the iteratively re-adjusting the at least one of the plurality of sentiment parameters comprises performing reinforcement learning based on the at least one reward point assigned to each of the plurality of robots, for each of the plurality of use cases.
10. A task execution device for task execution in dynamic heterogeneous robotic environment, the task execution device comprising:
a processor; and
a memory communicatively coupled to the processor, wherein the memory stores processor instructions, which, on execution, causes the processor to:
extract data associated with a plurality of data categories in response to identifying at least one current task, wherein the data is extracted from a plurality of data sources, and wherein each of the plurality of data categories is associated with at least one of the plurality of data sources;
derive a plurality of factors from the extracted data based on a deep learning network;
determine a plurality of correlations among the plurality of factors based on the deep learning network, wherein the plurality of correlations comprise weighted relationship between two or more factors from the plurality of factors;
derive a plurality of sentiment parameters for a set of factors from the plurality of factors based on the plurality of correlations, wherein the plurality of sentiment parameters are hierarchical;
simulate execution of the at least one current task by employing a plurality of robots based on each of the plurality of sentiment parameters applied to at least one associated factor from the set of factors;
iteratively re-adjust at least one of the plurality of sentiment parameters based on reinforcement learning performed on a result of the simulating, wherein the iterative re-adjustment is performed until each of the plurality of sentiment parameters are above a predefined threshold; and
execute the at least one current task by employing the plurality of robots, based on each of the plurality of sentiment parameters applied to at least one associated factor from the set of factors in response to the iterative re-adjusting.
11. The task execution device of claim 10 further comprising receiving the at least one current task requiring execution from a task pipeline.
12. The task execution device of claim 11 , wherein the extracting the data comprises:
determining a current state of the task execution device, in response to receiving the at least one current task; and
identifying the data based on the current state of the task execution device and the at least one current task received.
13. The task execution device of claim 10 , wherein the plurality of data sources comprises an environmental factors database and a historical database, wherein the historical database comprises at least one of order information, factory functioning information, external information, robot history information, or task handling information.
14. The task execution device of claim 10 , wherein the plurality of factors comprises at least one of date, day, time, product visibility, product character, or product category, and wherein each of the plurality of factors represents a neuron in the deep learning network.
15. The task execution device of claim 10 , wherein the determining the plurality of correlations among the plurality of factors comprises:
assigning weights to relations between a set of factors from the plurality of factors to generated weighted relationship between the set of factors;
analyzing the assigned weights based on a deep neural network; and
re-adjusting the assigned weights based on a result of the analyzing.
16. The task execution device of claim 10 , wherein the hierarchy associated with a sentiment parameter of the plurality of sentiment parameters corresponds to at least one of roles, task allocation to the plurality of robots, robot functioning, or product categories.
17. The task execution device of claim 10 , wherein the simulating execution of the at least one current task comprises:
determining a plurality of use cases for simulating execution of the at least one current task, wherein each of the plurality of use cases comprises application of each of the plurality of sentiment parameters to at least one associated factor from the set of factors; and
assigning, for each of the plurality of use cases, at least one reward point to each of the plurality of robots, based on simulation of the at least one current task.
18. The task execution device of claim 17 , wherein the iteratively re-adjusting the at least one of the plurality of sentiment parameters comprises performing reinforcement learning based on the at least one reward point assigned to each of the plurality of robots, for each of the plurality of use cases.
19. A non-transitory computer-readable storage medium for task execution in dynamic heterogeneous robotic environment, having stored thereon, a set of computer-executable instructions causing a computer comprising one or more processors to perform steps comprising:
extracting data associated with a plurality of data categories in response to identifying at least one current task, wherein the data is extracted from a plurality of data sources, and wherein each of the plurality of data categories is associated with at least one of the plurality of data sources;
deriving a plurality of factors from the extracted data based on a deep learning network;
determining a plurality of correlations among the plurality of factors based on the deep learning network, wherein the plurality of correlations comprise weighted relationship between two or more factors from the plurality of factors;
deriving a plurality of sentiment parameters for a set of factors from the plurality of factors based on the plurality of correlations, wherein the plurality of sentiment parameters are hierarchical;
simulating execution of the at least one current task by employing a plurality of robots based on each of the plurality of sentiment parameters applied to at least one associated factor from the set of factors;
iteratively re-adjusting at least one of the plurality of sentiment parameters based on reinforcement learning performed on a result of the simulating, wherein the iterative re-adjustment is performed until each of the plurality of sentiment parameters are above a predefined threshold; and
executing the at least one current task by employing the plurality of robots, based on each of the plurality of sentiment parameters applied to at least one associated factor from the set of factors in response to the iterative re-adjusting.