Methods and systems for intelligent agricultural management
AssigneeThe Board of Trustees of the University of Illinois
InventorsNaira Hovakimyan, Pan Zhao, Nicolas Martin, Ran Tao, Jing Wu
Life of this patent
Estimating the term…
Abstract
Embodiments of the present disclosure may include a method for agricultural management, including receiving a trained first management policy that was trained with state information. Embodiments may also include training a second management policy using imitation learning. In some embodiments, the imitation learning uses the trained first management policy and partial state information in order to output action information, the partial state information being a portion of the state information.
Claims18
A method of using imitation learning to apply an amount of nitrogen (N) input to crops, comprising:
- training a second management policy based on an imitation learning method and a second set of state information to provide a trained second management policy, wherein the imitation learning method is based on a trained first management policy, wherein the trained first management policy was trained using a reinforcement learning method and a first set of state information, wherein the second set of state information comprises a subset of the first set of state information;
- receiving, at runtime, from at least one sensor, information indicative of at least one environmental condition;
- determining the amount of N input using the trained second management policy and the at least one environmental condition, wherein the amount of N input optimizes crop yield action and reduces environmental impact;
- and
- transmitting, to an automated system, the amount of N input, wherein transmitting the amount of N input causes the automated system to apply the amount of N input.
A system of using imitation learning to apply an amount of nitrogen (N) input to crops, comprising:
- one or more sensors configured to collect information indicative of at least one environmental condition;
- a controller having at least one processor and a memory configured to store program instructions, wherein the processor is operable to execute the program instructions to carry out operations, the operations comprising:
- training, by the controller, a second management policy based on an imitation learning method and a second set of state information to provide a trained second management policy, wherein the imitation learning method is based on a trained first management policy, wherein the trained first management policy was trained using a reinforcement learning method and a first set of state information, wherein the second set of state information comprises a subset of the first set of state information;
- receiving, at runtime, from the one or more sensors, information indicative of at least one environmental condition;
- determining the amount of N input using the trained second management policy and the at least one environmental condition, wherein the amount of N input optimizes crop yield action and reduces environmental impact;
- and
- transmitting, to an automated system, the amount of N input, wherein transmitting the amount of N input causes the automated system to apply the amount of N input.
A non-transitory computer-readable medium having stored thereon instructions that, when executed by a computing device, cause the computing device to:
- train a second management policy based on an imitation learning method and a second set of state information to provide a trained second management policy, wherein the imitation learning method is based on a trained first management policy, wherein the trained first management policy was trained using a reinforcement learning method and a first set of state information, wherein the second set of state information comprises a subset of the first set of state information;
- receive, at runtime, from at least one sensor, information indicative of at least one environmental condition;
- determine the amount of N input using the trained second management policy and the at least one environmental condition, wherein the amount of N input optimizes crop yield action and reduces environmental impact;
- and
- transmit, to an automated system, the amount of N input, wherein transmitting the amount of N input causes the automated system to apply the amount of N input.
Dependencies are inferred from the claim text.