System and method for controlling a robotic manipulator based on hierarchical reinforcement learning of control policies

A feedback controller for controlling a robotic manipulator to perform a task, comprising:

• a user interface configured to collect a feedback signal indicative of measurements of a state of the robotic manipulator performing the task;
• a memory configured to store a hierarchical reinforcement learning (HRL) neural network including (i) a nominal control policy trained to generate, based on the feedback signal, a nominal action transitioning the robotic manipulator in a precondition state region to a termination state region of the task and (ii) a recovery control policy trained to generate, based on at least a portion of the feedback signal, a recovery action transitioning the robotic manipulator in a failure state region to the precondition state region of the nominal control policy, wherein the recovery control policy is trained based on a recover policy reward to generate a switch-to-nominal action to transfer control to the nominal control policy to perform the task, wherein the recovery policy reward for training the recovery control policy is dependent on a nominal policy reward for training the nominal control policy;
• and
• a processor configured to:
• iteratively execute the HRL neural network to generate an action from at least one of:
• the nominal action or the recovery action;
• and
• cause to control the robotic manipulator to perform the task based on the generated action.