P i (x i ;z i ) P o (x o ;z o ) P r (x r ;z r ) Remote site (wider FOV) Operator's site (narrower FOV) Scaling direction

Save this PDF as:
 WORD  PNG  TXT  JPG

Μέγεθος: px
Εμφάνιση ξεκινά από τη σελίδα:

Download "P i (x i ;z i ) P o (x o ;z o ) P r (x r ;z r ) Remote site (wider FOV) Operator's site (narrower FOV) Scaling direction"

Transcript

1 TVRSJ Vol.7 No.1, 22 HMD Dynamic Eects of Inconsistent Field of View in HMD-based Telexistence Systems Yasuyuki Yanagida 31 Susumu Tachi 32 Abstract { When using telexistence or virtual reality systems, the operator perceives the remote/virtual environment as a strange world, if the visual parameters of the remote camera or the projection used to generate the graphics images of virtual environments are inconsistent with those of the display device. For the systems using HMD, eld of view of the HMD (camera) and the inter-pupillary (inter-camera) distance can be considered as elementary parameters to determine the appearance of the displayed world. In this paper, this strangeness, which has been reported as subjective impression, is analyzed geometrically as a dynamic distortion and deformation of the perceived world. Also, a method to reduce the strangeness is proposed, by controlling the scaling factor of the rotational motion of the robot head so that the optical ow nearby the gazed point could match with the mechanism of vestibulo-ocular reex of human operators. Keywords : telexistence, virtual environment, head-mounted display, eld of view, vestibulo-ocular reex 1. (VR) [1] [2] VR (Head-Mounted Display: HMD) 1968 Sutherland [3] VR HMD *1 ATR *2 *1 The University oftokyo; currently at ATR Media Information Science Laboratories *2 The University oftokyo [4] HMD HMD [4] [5] [6] (CG) VR CG HMD [7] [8] [9] [1]

2 Vol.7, No.1, 22 HMD 2. HMD [11] HMD [12] [7] [13] [14] [15] [9] HMD HMD HMD VR CG [16] [17] [18] [19] (Augmented Reality) [2], (Mixed Reality) [21] CRT [22] HMD 2 HMD HMD 2 CG VR

3 : HMD = tan r tan o = tan o tan o (2) d d r =d o (3) 1 HMD Fig. 1 Geometric parameters for HMD-based telexistence system: inter pupillary distance (IPD)and eld of view (FOV) HMD HMD 1 HMD 1% 2 1 r o R L r, o, d r, d o r = o (1) O x z 3.2 P r (x r ;z r ) Rr, Lr tan Rr = x r+d r =2 z r (4) tan Lr = x rd r=2 z r (5) 1= tan Ro = 1 tan Rr = 1 1 x r+d r=2 z r (6) tan Lo = 1 tan Lr = 1 1 x rd r =2 z r (7) Ro, Lo P o (x o ;z o ) z = xd o=2 tan Ro z = x+d o=2 tan Lo (8) (9) P o x x o = d o d r x r = x r d (1) (9) z z o = d z r (11) P r (x r ;z r )

4 Vol.7, No.1, 22 1/ρ d Uniform scaling Changing IPD ρ ϕ Scaling in depth Changing FOV 2 Fig. 2 Static distortion due to inconsistent IPD and FOV Remote site (wider FOV) P r (x r, z r ) Head rotation θ P r (x r, z r ) θ Operators site (narrower FOV) P i (x i, z i ) Scaling direction Flow of the object in the world P o (x o, z o ) Distortion of the object Scaling direction 3 Fig. 3 Dynamic deformation of the world by head rotation when inconsistent eld of view is applied P o (x o ;z o )= xr ; z r = 1 (x r ; z r ) (12) d d d HMD [23] 3 HMD 3 P r (x r ;z r ) P i (x i ;z i ) P o (x o ;z o ) P r P i d =1 1 P i = P r (13) P r P o cos sin 1 P o = sin cos cos sin P r (14) sin cos

5 xo z o 2 = 6 4 : HMD x r (cos 2 + sin 2 3 ) +z r (1 ) sin cos 7 x r (1 ) sin cos 5 +z r (sin 2 + cos 2 ) (15)! xo xr + z r (1 ) z o x r (1 ) + z r xr zr = +(1 ) z r x r xi zi = = +(1 ) (16) z i x i 1 P i 2! _ OP o =(1 ) zo = d x o dt (17) P i P o P o HMD ( j=1)!! _ OP o OP o 1 j! 2 1 OP o j = 1 x o z o p 1 d x 2 o + zo 2 dt (18) (x o =) (z o ;x o ) xo z 2 o p x 2 + o z2 o 1 d dt (19) jx o jjz o j (19) > 1 d=dt < 1 (19) d=dt 4 HMD =2 3 [deg.] Remote Site Operator s Site 4 Fig. 4 Simulation result for dynamic deformation and ow of objects caused by inconsistent FOV HMD HMD (VOR: vestibulo-ocular reex) simulator sickness [24] VR 5 (a) HMD

6 Vol.7, No.1, 22 ϕ dθ (a) Normal case βϕ Original position /θ (c) FOV * β Actual position Actual position ϕ αdθ Original position (b) Rotation * α βϕ α/θ (d) Rotation * α & FOV * β (VOR conformable) 5 Fig. 5 Flow of the image on the screen when FOV and/or head rotation scaled (b) >1 (c) HMD >1 HMD (d) VOR > (14) cos sin 1 P o = sin cos cos sin P r (2) sin cos x r (cos cos + sin sin ) xo +z r (cos sin sin cos ) = x r (sin cos cos sin ) 5 z o +z r (sin sin + cos cos ) (21)! xo xr + z r ( ) z o x r (1 ) + z r xr ( )z r = + z r (1 )x r xi (= 1)z i = + (22) z i (1 )x i! _ (= 1)z o d OP o = (1 )x o dt (23) 3. 3 P o OP o (= 1)z o (1 )x o = x o z o (24)

7 : HMD P i (x i, z i ) P r (x r, z r ) Scaling direction Head rotation P o (x o,z o ) Stabilizing lateral image flow in this region P r (x r, z r ) Scaling direction αθ θ Remote Site Operator s Site Remote site (wider FOV) Rotation scaling Operators site (narrower FOV) 6 Fig. 6 Stabilization of the image by scaling rotational motion o x o =z o = 1 1 = 2 o (25) = 1+ 2 o o (26) o o = = (27) (27) (23)! _ d OP o = (28) (1 2 )x o dt (x ) x o = z / 7 7 ( ) Fig. 7 Stabilized world by rotation scaling (simulation result). 4 =2 3 [deg.] 15 [deg.] (= 3= ) [5] ( 8) 6.25 [deg] 1 [mm] 7 (NEC PC-981BX, CPU i486sx 2MHz + ODP)

8 Vol.7, No.1, 22 8 Fig. 8 Telexistence master system used for experiment (, 2 [Mbps]) 2 [khz] VR Silicon Graphics Indigo2 High IMPACT (CPU MIPS R44 2MHz) (IMPACT Channel Option) 1 PC ISA OpenGL [25] NTSC 6 [Hz] NTSC RGB 2 1/3 [s] HMD [5] 4 [deg.] TFT 5.2 1[m] 1[m] 1[m] 5[m] 4 (a) HMD ( =1) ( =1) (b) HMD ( =1) 2.3 ( =2:3) 2.3 HMD 4 [deg.] 8 [deg.] ( =2) (27) (c) HMD 2 ( =2) ( =1) (d) VOR 8 [deg.] ( =2) 2.3 ( =2:3) (d) (a) yaw

9 : HMD Angle (deg.) Angle (deg.) Time (sec.) (a)normal Target Head Time (sec) (c)fov scaled Target Head Angle (deg.) Angle (deg.) Time (sec.) Target Head (b)rotation scaled Time (sec.) Target Head (d)both roation and FOV scaled 9 Fig. 9 Head motion for target tracking. (b) (c) VOR (d) x(t), y(t) x(t)y(t + ) xy () = q q x 2 (t) y 2 (t) (29) 1 VOR 2 1 Cross Correlation Cross Correlation 1 (a).9 (b) (c).8 (d) Time Lag (sec) (a)yaw axis (a) (b) (c) (d) Time Lag (sec) (b)pitch axis 1 Fig. 1 Cross-correlation between target and head motion 1 (Max) (Lag) Table 1 Maximal value (Max)and its time lag (Lag)of cross correlation. Yaw Pitch Max Lag [s] Max Lag [s] (a) (b) (c) (d) VOR 6. HMD / simulator sickness

10 Vol.7, No.1, 22 HMD [1], :,, [2] :,, [3] I. E. Sutherland: A Head-Mounted Three- Dimensional Display, Proc. Fall Joint Computer Conference, AFIPS Conf. Proc., Vol. 33, pp. 757{ 764, [4], :,, Vol. 7, No. 4, pp. 314{326, [5],, :,, Vol. 1, No. 5, pp. 655{665, [6], :,, Vol. 115-C, No. 2, pp. 172{181, [7] W. Robinett and J. P. Rolland: A Computational Model for the Stereoscopic Optics of a Head- Mounted Display, Presence, Vol. 1, No. 1, pp. 45{ 62, [8] L. F. Hodges and E. T. Davis: Geometric Considerations for Stereoscopic Virtual Environments, Presence, Vol. 2, No. 1, pp. 34{43, [9] J. P. Rolland and W. Gibson: Towards Quantifying Depth and Size Perception in Virtual Environments, Presence, Vol. 4, No. 1, pp. 24{49, [1] :,, [11] :,, Vol. 444, pp. 25{223, [12],, : HMD: 3DDAC Mk. 4,, Vol. 4, No. 1, pp. 261{268, [13], : HMD,, Vol. 12, No. 6, pp. 911{918, [14],, : HMD,, Vol. 13, No. 4, pp. 58{582, [15], Paul Milgram,, :, Human Interface News & Report, Vol. 9, No. 3, pp. 295{32, [16] M. R. Mine: Characterization of End-to-End Delays in Head-Mounted Display Systems, Tech. Rep. TR93-1, Dept. of Computer Science, University of North Carolina at Chapel Hill, [17],,, : HMD, 6, pp. 1{8, 199. [18], :,, Vol. 115-C, No. 2, pp. 32{31, [19],, : HMD,, Vol. 4, No. 4, pp. 589{598, [2] R. Azuma: A Survey of Augmented Reality, Presence, Vol. 6, No. 4, pp. 355{385, [21] Y. Ohta and H. Tamura ed.: Mixed Reality Merging Real and Virtual Worlds, Ohmsha, [22] D. B. Diner and D. H. Fender: Human Engineering in Stereoscopic Viewing Devices, Plenum Publishing Corporation, [23], :, Human Interface News & Report, Vol. 9, No. 4, pp. 467{472, [24] : VE,, Vol. 3, No. 2, pp. 31{39, [25], :,,Vol. 115-C, No. 2, pp. 236{244, ( ) ( ) (ATR) ( )