Interpret Goal States in Motor System & Switch Policies

Enable the motor system to switch between different ways to achieve goals.

We would like to implement a state-switching mechanism where the motor system can switch the policy that it is executing, depending on the goal state it receives.

For example, we might want to do a semi-random walk when the motor system is not receiving goal states. We may want to switch to a saccade policy (rotating the sensor without moving it's position) when the sensor module sends a goal based on detected salient features. We may want to switch to a "JumpToGoalState" policy when the motor system receives a goal from the learning module that includes both a desired location and orientation (i.e. move the agent and rotate it)

This task also relates to Enable Switching Between Learning and Inference-Focused Policies.

---

Help Us Make This Page Better

All our docs are open-source. If something is wrong or unclear, submit a PR to fix it!

Make a Contribution

Learn how to contribute to our docs