Consider a way to define logic to override position closed-loop

This would work by overriding how the treadmill->vr position is implemented. It would be nice to expose the following features:
- Ability to define an active gain per site
- Ability to define a passive velocity per site

Final velocity would thus be given by:
`final_velocity = treadmill_velocity x passive_gain + active_gain`

One would need to decide how these transformations interact with downstream logic such as stopping logic.