Presented by Byungjoo Lee at CHI'16 San Jose
ABSTRACT
The computer mouse is rarely used for drawing due to its body-fixed coordinate system, which creates a stroke that differs from the user’s original hand movement. In this study, we resolve this problem by implementing a new mouse called StereoMouse, which eliminates the rotational disturbance of the coordinate system in real-time. StereoMouse is a special mouse with two optical sensors, and its coordinate orientation at the beginning of a stroke is maintained throughout the movement by measuring and compensating for the angular deviation estimated from those sensors. The drawing performance of StereoMouse was measured by means of having users perform the task of repeatedly drawing a basic shape. The results of this experiment showed that StereoMouse eliminated the horizontal drift typically observed in a stroke drawn by a normal mouse. Consequently, StereoMouse allowed the users to draw shapes at a 10.6% faster mean speed with a 10.4% shorter travel time than a normal mouse would. Furthermore, StereoMouse showed 37.1% lower chance of making incorrect gesture input than the normal mouse.
CHI'16 Journal "A Mouse With Two Optical Sensors That Eliminates Coordinate Disturbance During Skilled Strokes"
1. A Mouse with Two Optical Sensors that Eliminates
Coordinate Disturbance During Skilled Strokes
Human–Computer Interaction 30(2) 2015
Byungjoo Lee and Hyunwoo Bang
Seoul National UniversityFirst prototype
Second prototype
Wooden frame
3D printed assembly
2. Contents
1. Decade-old Problem of Computer Mouse
2. Modelling and the Solution
3. Implementation of StereoMouse
4. User Study
5. Conclusion
5. • 15% longer time, 50% larger error at the same time.
Zabramski et al., 2011
Error plotTime plotShape tracing task
Problem of Computer Mouse
6. • Out of Path Movement (OPM) for mouse was 97%
higher than for stylus.
Sergey et al., 2005
Problem of Computer Mouse
7. • Aweful subjective ratings…
Greenberg et al., 1992
Problem of Computer Mouse
8. However Not Bad for Pointing
• Mouse shows very efficient pointing or dragging
performance.
0bit/s
1,25bit/s
2,5bit/s
3,75bit/s
5bit/s
Pointing Dragging
Mouse Stylus w/ tablet
Makenzie et al., 1991
Cockburn et al., 2012
9. Mouse vs. Stylus
Mouse Stylus w/ tablet
Price Low High
Portability Good Bad
Pointing performance Good Good
Steering/drawing
performance
Bad Good
Maintenance Easy Hard
Workspace Broad Limited
10. Objective of This Research
StereoMouse Stylus w/ tablet
Price Low High
Portability Good Bad
Pointing performance Good Good
Steering/drawing
performance
Adaptable Good
Maintenance Easy Hard
Workspace Broad Limited
15. • The rotation of the wrist joint was naturally transformed
into a straight line in the display space.
Result
Input
Input Result
Mouse Movement ≠ Pointer Movement
17. (1) How much the mouse will rotate?
(2) How the optical sensor will transfer the user
movement?
Modelling Coordinate Disturbance
18. • A redundant 2D manipulator with 3 revolute joints.
−1 −0.5 0 0.5 1 1.5 2
0
1
2
3
4
5
6
7
8
Angle (rad)
Posturecost
Shoulder
Elbow
Wrist
e
s
w
Modelling Mouse Rotation
Elbow
Shoulder
Wrist
19. • Using parabolic cost functions for each of joint
Modelling Mouse Rotation
a 4. + + + 4.
1 2 3 4 5
+ + + + + +
6 7 8 9 10 11
+ + + + 4- 4- 4-
A 13 14 15 16 17 is
~ 21 ~9 23
4- 4- 4- 4- 4- 4- 4-
28 29 30 31 32 33 34
b
4- +
36 37
+ +
38 39
~.. +
4O
Fig. 1. a Top view of the experimental arrangement. The 41 target
points are placed on a horizontal plane in front of the subject. The
three measured angles are marked, b A schematical drawing of
. or/
20 ,'*0 60 80 100 120 140 160
elbow angle {deg)
Cl
9 ~,
, ,
-60 --'-0 -20 0 20 40 60 80
shoulder angte [de<J}
5
ml
a~'~,a-_ _ j l
0 i , i i i i ~r-r-i i i i i ~ 1 i
20 40 60 80 100 120 140 160
elbow angle (deg)
b
1~15
u
9 , , , , i i , i i ~ - i i i i , !
-60 -40 -20 O 20 40 60 80
shoulder angle (deg)
-~-~'~s~ -
"~ ~ ' I-~. _g =
20 ~0 60 SO
C
:1-60 -40 -20
s
-60 -:0 -2Q 0 20 /-0 60 80
wrist angle [de9)
_,- .j;,"5
9 i | w i i i~l i i i i i w i
29 L,O 60 80 100 120 140 160
elbow angle [deg)
d
:1 i - r | ~ i i i i i i i
-60 -40 -20 0 20 LO 60 80
shoulder angle (deg)
el
t '" "o . . . . . . . ~,'~,'~--~,'X,',", , ,
-so-~.o-20 o 20 ~,o 6o e 9
wrist anglo (deg}
9--" _,,-
0 I I I 1 i- i--i i I I l I I I l
20 /,0 60 80 109 120 14.0 160
elbow angle (deg)
e
i i i | i i i i F i i i
-60 -40 -20 0 20 40 60 80
shoulder angle (deg)
Fig. 3a-e. Psych 9
"comfort" functions fo
from five subjects a to
angle in degrees. Ordin
subjective discomfort i
the elbow those parabo
which give the best app
values. For shoulder a
+ S.D. range of those p
is shown which are obt
approximating the targ
H. Cruise et al., 1990
Elbow
Shoulder
Wrist
−1 −0.5 0 0.5 1 1.5 2
0
1
2
3
4
5
6
7
8
Angle (rad)
Posturecost
Shoulder
Elbow
Wrist
e
s
w
20. • Solving inverse kinematics with minimizing the cost
function:
Desk coordinate system
Region of interest
Neutral posture
Shoulder Elbow
Wrist
s =1.0
e =1.0
w = 0.25
( 0.5, 0) (0.5, 0)
(0, 0)
(0, 0.5)
(0, -0.5)
x
y
e0 = 80°
s0 = 0°
w0 = 0°
Horizontal deviation
Verticaldeviation
Amount of mouse rotation (deg)
−0.5 0 0.5
−0.5
0
0.5 −40
−20
0
20
40
(+) (-)
Amount of end effector rotation (deg)
{costw(ɵw)+ coste(ɵe)+ costs(ɵs)}
Horizontal deviation
Verticaldeviation
Modelling Mouse Rotation
21. O
s
Physical motion
Disturbed motion
Displacement of cursor
net
net
effective
r
r(1 cos net )
rsin net
r
Disturbance in orientation
Disturbance in length
Modelling Transfer Function of Mouse
•Transfer function within a time step
22. Verticaldeviation
Cursor movement Hand movement
Horizontal deviation
−1.5 −1 −0.5 0
−0.5
0
0.5
1
Horizontal deviation
Verticaldeviation
Cursor movement Hand movement
−0.8 −0.6 −0.4 −0.2 0
−0.4
−0.2
0
0.2
0.4
Verticaldeviation Cursor movement Hand movement
Horizontal deviation
−1.5 −1 −0.5 0
−0.5
0
0.5
1
Horizontal deviation
Verticaldeviation
Cursor movement
Hand movement
−0.8 −0.6 −0.4 −0.2 0
−0.4
−0.2
0
0.2
0.4
Cursor Movement Hand Movement
Simulated
Result of
Coordinate
Disturbance
23. Verticaldeviation
Cursor movement Hand movement
Horizontal deviation
−1.5 −1 −0.5 0
−0.5
0
0.5
1
Horizontal deviation
Verticaldeviation
Cursor movement Hand movement
−0.8 −0.6 −0.4 −0.2 0
−0.4
−0.2
0
0.2
0.4
Verticaldeviation Cursor movement Hand movement
Horizontal deviation
−1.5 −1 −0.5 0
−0.5
0
0.5
1
Horizontal deviation
Verticaldeviation
Cursor movement
Hand movement
−0.8 −0.6 −0.4 −0.2 0
−0.4
−0.2
0
0.2
0.4
Cursor Movement Hand Movement
Simulated
Result of
Coordinate
Disturbance
StereoMouse
compensates
the distortion
24. Basic Idea of StereoMouse
• Coordinate orientation is maintained throughout the
movement by compensating for the angular deviation.
θ =∫dθ
t=0
t=i
i
t=i
Linear velocity measured at i-th step
i
Before correction After correction
26. Hardware of StereoMouse
• StereoMouse has two optical sensors to measure its
body rotation.
2 x Arc Touch mouse | 2 x Wedge mouse
Sensor resolution: 1000 counts per inch
Sensor refresh rate: 125 Hz
Angular resolution:
= 0.0191 ~ 0.0286 degrees/count
First prototype
Second prototype
Wooden frame
3D printed assembly
First Prototype
Second Prototype
Sensor#1
Sensor#2
d
Sensor #2
Sensor #1
27. Software of StereoMouse
The data flow
Laptop computer
Sensor #1
Sensor #2
StereoMouse
Multipoint
mouse SDK
Java
application
Bluetooth
+
(dx1,dy1,button_ state)
(dx2,dy2,button_ state)
(d , )
(dx,dy)
(x, y)
28. Accuracy of Angular Measurement
Stepping motor
Motor controller
Microprocessor
Rotary table
Timing belt
Rotary table for the measurement
34. Two User Studies
1.Drawing without visual feedback (open-loop)
— to test the predictability of the model
2.Drawing with visual feedback (closed-loop)
— to measure kinematic performance after visual
correction
36. Experimental Design
• Drawing four primitive shapes (Circle, Triangle,
Rectangle, Star) x 10 times.
The experimental setup
37. Experimental Design: Dependent Measure
• The aperture between start and end points of a
drawing.
Aperture
Start point
End point
38. Result: Aperture
• Device mode on the aperture between the start and
end point (F(1,8)=16.44; p < .004).
• StereoMouse showed half of the aperture for the
normal mouse condition.
StereoMouseNormal mouse
Aperture(pixels)
4000
3000
2000
1000
0
StereoMouse
Normal mouse
42. Experimental Design
• Drawing four primitive shapes (Circle, Triangle,
Rectangle, Star) in two sizes (Large, Small) x 8 times.
The experimental setup
49. Take-away Notes
1. Mechanical problems in HCI are relatively
under-investigated.
2. Mechanistic modelling is a powerful method to
describe and solve the problems.