2011/09/19
Jay
son Ding
1
Literature review robot door open
1.1
Pulling open novel doors and drawers with equilibrium point control
(IROS2009)
Dr. Charles Kamp’s group empirically demonstrate that equilibrium point control can enable a
humanoid robot to pull open a variety of doors
and drawers without detailed prior models, and
infer their kinematics in the process.
[1]
Their implemen
tation uses a 7 DoF anthropomorphic
arm with series elastic actuators (SEAs) at each joint, a hook as an end effector, and low
mechanical impedance. The control algorithm and the model of robot hook to pull door are
showed in above figure.
1.1.1
Technical details
The Equilibrium Point Controller (EPC) can be expressed as:
Where
t
x
eq
is the equilibrium point in Cartesian coordinate ,
t
q
eq
is the equilibrium
point is
joint space.
t
m
and
t
h
are control vector for position and force controller.
They present two controllers

pull linear and pull radial force that generate a CEP at each time
step (ca. 100ms) and use inverse
kinematics to command the arm with the corresponding JEP.
The pull linear controller produces a linear CEP trajectory and is expressed as:
2011/09/19
Jay
son Ding
The pull radial force controller alters its CEP trajectory based on real

time estimates of the
mechanism’s kinem
atics and is expressed as:
For each SEA actuator, they applies a gravity compensating torque plus a torque from a
simulated, torsional, viscoelastic spring. Each virtual spring has constant stiffness and damping,
and a variable equilibrium angle. These
equilibrium angles form a joint space equilibrium point
(JEP), which has a corresponding Cartesian space equilibrium point (CEP) for the arm’s end
effector.
1.1.2
Results
In their paper, they
present results from empirical evaluations of their performance (108
trials).
In these trials, both controllers were robust with respect to variations in the mechanism, the
pose of the base, the stiffness of the arm, and the way the handle was hooked.
They
also tested
the more successful controller with 12 distinct mechanisms. In these tests, it was able to open
11 of the 12 mechanisms in a single trial, and successfully categorized the 11 mechanisms as
having a rotary or prismatic joint, and opening to the
right or left. Additionally, in the 7 out of 8
trials with rotary joints, the robot accurately estimated the location of the axis of rotation.
1.2
Pulling open doors and drawers: Coordinating an omni

directional base
and a compliant arm with Equilibrium Point
control
(ICRA2010)
Dr. Charles Kamp’s group also has studied of coordinating
omni

directional base
with robot arm
to open a door
[2]
.
With
this
new methods,
they
show that the robot can autonomously approach
and open doors
and drawers
for which only the location and orientation of the handle have been
provided.
2011/09/19
Jay
son Ding
1.2.1
Technical details
The controller is similar to their EPC door open controller. The difference is
the calculation of the
velocity driving vector
base
v
for mobile base which is calculated as:
s
cm
r
r
v
base
/
15
where
C
p
l
close
l
close
p
x
p
x
r
2
p
is the point of adjacent boundary of
cm
p
x
t
s
Bndry
p
C
l
close
10
.
.
and
l
close
x
is the
closest point to robot arm workspace boundary expressed as
Bndry
x
dis
x
l
ee
l
eq
x
x
x
l
close
,
min
arg
,
.
1.2.2
Results
They
also demonstrate that EPC can coordinate the movement of the robot's omni

directional
base and compliant arm while pulling open a d
oor or drawer, which leads to significantly
improved performance. Through 40 trials with 10 different doors and drawers,
they
empirically
demonstrated the robustness of the system. The robot succeeded in 37 out of 40 trials, and had
no more than a single f
ailure for any particular door or drawer.
1.3
Learning to Open New Door
(IROS 2010)
Dr. Andrew. NG’s group from Stanford University presents a visual guided approach to assist
robot door open
[3]
.
2011/09/19
Jay
son Ding
1.3.1
Technical detail
They use an image of the door handle to identify a small number of 3d key locations; such as
the axis of rotation of the door handle(V1), and the location of the end

point(P1) of the door

handle(P2) as s
howed in the above figure. These key locations then completely define a
trajectory for the robot end

effector (hand) that successfully turns the door handle and opens the
door.
Door handle 3d feature extraction:
The authors did not use the stereo camera
to extract the 3d key location, since that data is too
noisy and resulting in a large area of erroneous depth values in the center of handle. Hence
author used a set of 2d coordinate from camera image and translated them into 3d coordinate
using the laser
scanner. Here they also assume the door is always vertical to the ground. The
equation to estimate the distance between camera origin to the door handle can be formulated
as:
i
i
d
l
tv
T
t
2
2
0
*
min
Where
3
2
0
R
T
is the matrix which projects camera to laser plane,
i
l
is the door trajectory in
laser scanner plane,
d
is the horizontal distance between camera and laser,
3
R
v
is the
vector pointing from door h
andle to camera origin. A geometric relationship for those vector is
showed in the following figure.
2011/09/19
Jay
son Ding
After extracting the door handle 3d feature, they generated the trajectory for robot t
o open the
door. The 3d location of the robot arm can be expressed as:
*
1
r
T
p
Where
4
4
1
R
T
from the camera frame to robot frame
1.3.2
Results
Evaluated on a large set of doors that the robot had not previously seen, it successfu
lly opened
31 out of 34 doors. We also show that this approach of using vision to identify a small number of
key locations also generalizes to a range of other tasks, including turning a thermostat knob,
pulling open a drawer, and pushing elevator buttons.
1.4
Planning for autonomous door opening with a mobile manipulator
Willow garage cooperated with Grasp lab at University of Pennsylvania used
a graph

based
search algorithm to navigate Pr2 robot open a door
[4]
. They
graphically repre
sent robot door
open and use graphic searching algorithm which is small enough for efficient planning and rich
enough to contain feasible motions that open doors to solve robot door open problem.
1.4.1
Technical details
Graphic representation of the robot door open task. As showed in above middle figure, door
interval 0 contained all the door fully closed position and door interval 1 contained all the door
2011/09/19
Jay
son Ding
open position. A searching algorithm i
s doing path planning to guide the robot moving from
Interval 0 to interval 1. The path planning is only subjected to two constraints: a) the based
cannot collide with any part of the environment. b) the motion of the base must allow the gripper
to stay on
the handle so that the door can be pushed or pulled open by the arm. The cost
function is combined the 2d moving cost and the distance between robot handle to shoulder. A
most popular graphic searching algorithm
[5]
is used in these application.
1.4.2
Results
They demonstrate our approach on the PR2 robot

a mobile manipulator with an
omnidirectional base and a 7 degree of freedom arm. The robot was successful in opening a
variety of door
s both by pulling and pushing.
1.5
Bio

inspired assistive robotics: Service dogs as a model for human

robot
interaction and mobile manipulation
Other researchers proposed a way to re
engineer the environment in order to help assistive
robots perform
door open
w
ith generality and robustness.
Like Dr. Charles Kemp group
present
a biologically inspired robot capable of obeying many of the same commands and exploiting the
same environmental modifications as service dogs
[6]
. The robot responds to a subset of the 71
verbal commands listed in the service dog training manual used by Georgia Canines
for
Independence. In
his
implementation, the human directs the robot by giving a verbal command
and illuminating a task

relevant location with an off

the

shelf green laser pointer.
Figure
1
Robot simulated service to open the door.
1.5.1
Technical details
They used laser pointer to lead the robot close to the door and face directly to the tower. The
robot will used a color segmentation method to extra the tower. After segmenting the towel, the
ro
bot attempts to find the bottom tip of the towel. With this 2D coordinate we then calculate the
3D position of this pixel by assuming that the 2D towel pixel lies on the plane estimated by the
laser range finder which typically corresponds with the surface
of the door or drawer.
1.5.2
Results
In particular,
they
show that by tying or otherwise affixing colored towels to doors and drawers
an assistive robot can robustly open these doors and drawers in a manner similar to a service
dog. In our tests, the robot successfully opened two different drawers in 18 out of 2
0 trials (90%),
2011/09/19
Jay
son Ding
closed a drawer in 9 out of 10 trials (90%), and opened a door that required first operating a
handle and then pushing it open in 8 out of 10 trials (80%).
1.6
References
[1]
A. Jain and C. C. Kemp, "Pulling open novel doors a
nd drawers with equilibrium point
control," in
Humanoid Robots, 2009. Humanoids 2009. 9th IEEE

RAS International
Conference on
, 2009, pp. 498

505.
[2]
A. Jain and C. C. Kemp, "Pulling open doors and drawers: Coordinating an omni

directional base and a comp
liant arm with Equilibrium Point control," in
Robotics and
Automation (ICRA), 2010 IEEE International Conference on
, 2010, pp. 1807

1814.
[3]
E. Klingbeil
, et al.
, "Learning to open new doors," in
Intelligent Robots and Systems
(IROS), 2010 IEEE/RSJ Intern
ational Conference on
, 2010, pp. 2751

2757.
[4]
S. Chitta
, et al.
, "Planning for autonomous door opening with a mobile manipulator," in
Robotics and Automation (ICRA), 2010 IEEE International Conference on
, 2010, pp.
1799

1806.
[5]
P. E. Hart
, et al.
, "A F
ormal Basis for the Heuristic Determination of Minimum Cost
Paths,"
Systems Science and Cybernetics, IEEE Transactions on,
vol. 4, pp. 100

107,
1968.
[6]
H. Nguyen and C. C. Kemp, "Bio

inspired assistive robotics: Service dogs as a model for
human

robot in
teraction and mobile manipulation," in
Biomedical Robotics and
Biomechatronics, 2008. BioRob 2008. 2nd IEEE RAS & EMBS International Conference
on
, 2008, pp. 542

549.
Σχόλια 0
Συνδεθείτε για να κοινοποιήσετε σχόλιο