ML-Agents Imitation Learning and Oculus

Hi, can anyone suggest the best way to set up imitation learning where the Oculus avatar is the source of the inputs for agentAction?
I need the VR player to be imitated. Thanks, Mike

Oculus helped with this solution:

The OVRCameraRig (part of the OVRPlayerController prefab) is what controls the player view, and the OVRManager is what sends the input.

The headset can be controlled with HMD emulation.

Instead of mapping a button directly to an action, you’d map a button to a function which would then do that action. You could then call that function as needed (simulating a button press). Similarly, you can lock out a controller by checking for a boolean variable, and only calling the function if true.