Reinforcement learning based satellite control

A method for implementing a reinforcement learning model for controlling a satellite, the method comprising:

• receiving a set of parameters associated with an orientation of the satellite;
• and
• implementing, based on each of the set of parameters, the reinforcement learning model to control the orientation of the satellite to stabilize the satellite, wherein the implementing of the reinforcement learning model comprises:
• reducing, by an omega agent, a value of each of a set of angular rates such that each of the set of angular rates is within a first predefined threshold;
• reducing, by a Euler agent when each of the set of angular rates is within the first predefined threshold, a value of each of a set of Euler angles such that each of the set of Euler angles is within a second predefined threshold;
• predicting, based on each of the set of parameters, a set of actions to be implemented by the satellite to control the orientation of the satellite;
• implementing the set of actions;
• and
• predicting an outcome of the implementing of the set of actions to generate a set of rewards, wherein the orientation of the satellite is controllable, based on the set of actions and the set of rewards, to stabilize the satellite.

A method for implementing a reinforcement learning model for controlling a satellite, the method comprising:

• receiving a set of parameters associated with an orientation of the satellite;
• and
• implementing, based on each of the set of parameters, the reinforcement learning model to control the orientation of the satellite to stabilize the satellite, wherein the implementing of the reinforcement learning model comprises:
• predicting, based on each of the set of parameters, a set of actions to be implemented by the satellite to control the orientation of the satellite;
• implementing the set of actions;
• and
• predicting an outcome of the implementing of the set of actions to generate a set of rewards, wherein the orientation of the satellite is controllable, based on the set of actions and the set of rewards, to stabilize the satellite;
• capturing, by a simulation environment processor, a plurality of trajectories for the satellite during different orientation scenarios in different simulation environments, wherein each trajectory comprises data that represents a particular orientation of the satellite at a particular time;
• retrieving, by the reinforcement learning model, a set of trajectories of the satellite;
• and
• processing, by the reinforcement learning model, the set of trajectories of the satellite to learn and generate the set of parameters that describe the orientation of the satellite, wherein the set of parameters are processible by an agent to generate the set of actions for controlling the satellite.

A system for controlling a satellite, the system comprising:

• at least one hardware-based processor and memory, wherein the memory comprises processor-executable instructions encoded on a non-transient processor-readable media, wherein the processor-executable instructions, when executed by the processor, configure the system to:
• receive a set of parameters associated with an orientation of the satellite;
• execute, based on each of the set of parameters, a reinforcement learning model to control the orientation of the satellite to stabilize the satellite, wherein the executing of the reinforcement learning model comprises:
• reduce, by an omega agent, a value of each of a set of angular rates such that each of the set of angular rates is within a first predefined threshold;
• reduce, by a Euler agent when each of the set of angular rates is within the first predefined threshold, a value of each of a set of Euler angles such that each of the set of Euler angles is within a second predefined threshold;
• predict, based on each of the set of parameters, a set of actions to be executed by the satellite to control the orientation of the satellite;
• execute the set of actions;
• and
• predict an outcome of the executing of the set of actions to generate a set of rewards;
• and
• control, based on the set of actions and the set of rewards, the orientation of the satellite to stabilize the satellite.

A system for implementing a reinforcement learning model for controlling a satellite, the system comprising:

• at least one hardware-based processor and memory, wherein the memory comprises processor-executable instructions encoded on a non-transient processor-readable media, wherein the processor-executable instructions, when executed by the processor, are configurable to configure the system to:
• receive a set of parameters associated with an orientation of the satellite;
• and
• implement, based on each of the set of parameters, the reinforcement learning model to control the orientation of the satellite to stabilize the satellite, wherein the implementing of the reinforcement learning model comprises:
• reduce, by an omega agent, a value of each of a set of angular rates such that each of the set of angular rates is within a first predefined threshold;
• reduce, by a Euler agent when each of the set of angular rates is within the first predefined threshold, a value of each of a set of Euler angles such that each of the set of Euler angles is within a second predefined threshold;
• predict, based on each of the set of parameters, a set of actions to be implemented by the satellite to control the orientation of the satellite;
• implement the set of actions;
• and
• predict an outcome of the implementing of the set of actions to generate a set of rewards, wherein the orientation of the satellite is controllable, based on the set of actions and the set of rewards, to stabilize the satellite.

A non-transitory computer-readable medium storing a set of instructions for implementing a reinforcement learning model for controlling a satellite, the set of instructions comprising:

• one or more instructions that, when executed by one or more processors of a device, cause the device to:
• receive a set of parameters associated with an orientation of the satellite;
• and
• implement, based on each of the set of parameters, the reinforcement learning model to control the orientation of the satellite to stabilize the satellite, wherein the implementing of the reinforcement learning model comprises:
• reduce, by an omega agent, a value of each of a set of angular rates such that each of the set of angular rates is within a first predefined threshold;
• reduce, by a Euler agent when each of the set of angular rates is within the first predefined threshold, a value of each of a set of Euler angles such that each of the set of Euler angles is within a second predefined threshold;
• predict, based on each of the set of parameters, a set of actions to be implemented by the satellite to control the orientation of the satellite;
• implement the set of actions;
• and
• predict an outcome of the implementing of the set of actions to generate a set of rewards, wherein the orientation of the satellite is controllable, based on the set of actions and the set of rewards, to stabilize the satellite.

A non-transitory computer-readable medium storing a set of instructions for controlling a satellite, the set of instructions comprising:

• one or more instructions that, when executed by one or more processors of a device, cause the device to:
• receive a set of parameters associated with an orientation of the satellite;
• execute, based on each of the set of parameters, a reinforcement learning model to control the orientation of the satellite to stabilize the satellite, wherein the executing of the reinforcement learning model comprises:
• reduce, by an omega agent, a value of each of a set of angular rates such that each of the set of angular rates is within a first predefined threshold;
• reduce, by a Euler agent when each of the set of angular rates is within the first predefined threshold, a value of each of a set of Euler angles such that each of the set of Euler angles is within a second predefined threshold;
• predict, based on each of the set of parameters, a set of actions to be executed by the satellite to control the orientation of the satellite;
• execute the set of actions;
• and
• predict an outcome of the executing of the set of actions to generate a set of rewards;
• and
• control, based on the set of actions and the set of rewards, the orientation of the satellite to stabilize the satellite.