Refresh our human avatar

[severin-lemaignan]

This meta-issue tracks the progresses towards a more up-to-the-task human avatar

The 'high-level' aims are:

• clean, lightweight, good-looking model, with good-looking animations
• no more human-specific sensors/actuators (at least for the main features) -> use of regular armatures, motion actuators, etc
• cleaner code, less depend on Blender internal, to implement the human special features like the first person shooter perspective.

  Tasks

• [x] 3D model
  • [x] clean meshes build upon MakeHuman models (~3700 faces - still quite large)
  • [x] skinning in the BGE
  • [x] nice textures (thanks MakeHuman) -- caucasian male model for now
• [x] walk:
  • [x] basic walk animation implemented, corresponding Blender's 'action': walk
  • [x] add motion actuators to the human (at first, a keyboard actuator, then a waypoint actuator, then...) to move it, and to start/stop the walk cycle
• [ ] gestures:
  • [x] IK for arms, legs, head works reasonably well
  • [ ] make it easy for the human to point at something
  • [ ] re-enable and test grasping of objects (see also #569)
• [ ] create new animations
  • [ ] sitting
  • [ ] standing
• [ ] first person perspective (if needed!):
  • [ ] add a 'head' camera to see what the human sees
  • [ ] port the previous user interface (using keys + mouse) to create a real 'first person' user experience where we fully control the human, including pick and place tasks.
• [x] interface to external skeleton capture devices like kinect -> non-issue as the model rely on a standard armature. Issue #653 tracks the development of external scripts to control the avatar via dedicated peripherals like the kinect

  Related issues

• User report on various bugs with the human: #589
• generic integration with MakeHuman: #503 (this work has stalled. Needs to be re-taken over!)

[severin-lemaignan]

A first big step has been completed and merged: #682