[farshizzo]

The following script will load an avatar and display the skeletal structure of its bones and the local coordinate system of each bone (r=x,g=y,b=z).

import viz
viz.go()

def CreateSkeleton(avatar):
	
	AXIS_SIZE = 0.03
	viz.startlayer(viz.LINES)
	viz.linewidth(2)
	viz.vertexcolor(viz.RED)
	viz.vertex(0,0,0)
	viz.vertex(AXIS_SIZE,0,0)
	viz.vertexcolor(viz.GREEN)
	viz.vertex(0,0,0)
	viz.vertex(0,AXIS_SIZE,0)
	viz.vertexcolor(viz.BLUE)
	viz.vertex(0,0,0)
	viz.vertex(0,0,AXIS_SIZE)
	axes = viz.endlayer()
	
	points = []
	lines = []

	def traverse(bone):
		points.append(bone)
		for b in bone.getChildren():
			lines.append((bone,b))
			traverse(b)
			
	#Collect list of bones for drawing points and lines
	for b in avatar.getRootBoneList():
		traverse(b)

	#Create OTF object
	viz.startlayer(viz.LINES)
	viz.linewidth(1)
	viz.pointsize(6)
	viz.vertexcolor(viz.WHITE)

	for l in lines:
		viz.vertex(0,0,0)
		viz.vertex(0,0,0)

	viz.startlayer(viz.POINTS)

	for p in points:
		viz.vertex(0,0,0)

	skeleton = viz.endlayer()
	skeleton.dynamic()

	#Link each vertex to corresponding bone
	count = 0
	for b1,b2 in lines:
		viz.link(b1,skeleton.Vertex(count))
		count += 1
		viz.link(b2,skeleton.Vertex(count))
		count += 1

	for b in points:
		viz.link(b,skeleton.Vertex(count))
		count += 1
		viz.link(b,axes.clone(),srcFlag=viz.ABS_GLOBAL)

	axes.remove()
	
	return skeleton
	
model = viz.add('vcc_male.cfg',alpha=0.7)

CreateSkeleton(model)

import vizcam
cam = vizcam.PivotNavigate(distance=3)
cam.centerNode(model)

#Toggle avatar mesh on space key
vizact.onkeydown(' ',model.visible,viz.TOGGLE)

I believe the actual problem is that you are not applying an offset rotation when transferring the sensor data to the avatar. If you are using viz.AVATAR_WORLD coordinates to apply the data, then keep in mind that the initial pose of the avatar is the zeroed rotation of the hand. If the zeroed rotation of your hand sensor is pointing in a different direction, then you will need to compute the rotation difference between the two and apply that offset when transferring the data.

If you are using link objects, then you will need to apply a preEuler operator to the link to get the data to match the avatar.