Non-modal layout

[mayacakmak]

Here is a link to our low-fidelity prototypes of the interface changes we proposed to the group.

One of the key inefficiencies we observed in our earlier studies was the need for frequent mode switching. Since Henry is using the interface on a larger screen and the camera video stream does not fill this screen, we are envisioning having the different modes available all at once on one screen. We might revert this at some point (i.e. there might be some switches of mode) but it will be useful to implement as an option and will facilitate future features that require major layout configuration (e.g. adding new camera views to the screen).

In this layout there are two camera views corresponding to the navigation ("Drive" tab) and manipulation modes ("Arm" and "Hand" tabs). "Look" tab is eliminated by placing large buttons around the view for looking. "Arm" and "Hand" tabs are combined.

[mayacakmak]

I've made a lot of progress on this, while also doing a substantial refactoring (Issue #9). This includes:

• Two views rendering the camera stream for navigation and manipulation, each with its own overlay control
• Click to activate a view/mode (inactive is faded out and the controls are hidden)
• Manipulation overlay updated to include wrist, with griper control through outside buttons. Arrow changed for in/out to distinguish from up/down.
• Follow gripper checkbox

Still todo:

• [x] Move camera automatically for each view change
• [x] Add button to recenter view after using look buttons
• [x] (Not urgent) Look buttons on manipulation mode should add an offset to follow gripper (keep following with same offset)