Vous êtes sur la page 1sur 11

JOURNAL

IEEE OF ROBOTICS
AUTOMATION,
AND VOL.
NO. RA-3, FEBRUARY 1987 43

A Unified Approach for Motion and Force Control


of Robot Manipulators: The Operational Space
Formulation

Abstract-A framework for the analysis and control of manipulator forces. The magnitude of these dynamic forces cannot be
systems with respect to the dynamic behavior of their end-effectors is ignored when large accelerations and fast motions are consid-
developed. First, issues related to the description of end-effector tasks ered. Controlling the end-effector contact forces in some
that involve constrained motion and active force control are discussed.
Thefundamentals of the operational space formulation are thenpre- direction can be strongly affected by the forces of coupling
sented, andthe unified approach for motion and force control is created by the end-effector motion that can take place in the
developed. The extension of this formulation to redundant manipulator subspace orthogonal to that direction. The description of the
systems is also presented,constructing the end-effector equations of dynamic interaction between end-effector motions and the
motion and describing their behavior with respect to joint forces. These effects of these motions on the end-effector’s behavior in the
results are used in thedevelopment of a new and systematic approachfor
dealing with the problems arising at kinematic singularities.
At a singular direction of force control are basic requirements for the
configuration, the manipulator is treated as a mechanismthat is analysis and design of high-performance manipulator control
redundant with respect to the motion of the end-effector in the subspace systems. Obviously, these characteristics cannot be found in
of operational space orthogonal to the singular direction. the manipulator joint space dynamic model, whichonly
provides a description of the interaction between joint mo-
I. INTRODUCTION tions. High-performance control of end-effector motion and
contact forces requires the description of how motions along
R ESEARCH in dynamics of robot mechanisms has largely
focused on developing the equations of joint motions.
These joint space dynamic models have been the basis for
different axes are interacting, andhow the apparent or
equivalent inertia or mass of the end-effector varies with
configurations and directions.
various approaches to dynamic control of .manipulators.
The description, analysis, and control of manipulator sys-
However, task specification for motion and contact forces,
tems with respect to the dynamic characteristics of their end-
dynamics, and force sensing feedback are closely linked to the
effectors has been the basic motivation in the research and
end-effector. The dynamic behavior of the end-effector is one
development of the operational space formulation. The end-
of the most significant characteristics in evaluating the
effector equations of motion 131, 141 are a fundamental tool
performance of robot manipulator systems. The problem of
for the analysis, control, and dynamic characterization 181 of
end-effector motion control has been investigated, and al-
manipulator systems. In this paper, we will discuss, from the
gorithms resolving end-effector accelerations have been devel-
perspective of end-effector control, the issue of task descrip-
oped V I , [111, 1221, [3Ol, 1331.
tion, where constrained motions and contact forces are
The issue of end-effector dynamic modeling and control is
involved. The fundamentals of the operational space formula-
yet more acute for tasks that involve combined motion and
contact forces of the end-effector. Precise control of applied tion are presented, and the unified approach for the control of
end-effector forces is crucial to accomplishing advanced robot end-effector motion and contact forces is developed.
Treated within the framework of joint space control
assembly tasks. This is reflected by the research effort that has
been devoted to the study of manipulator force control. systems, redundahcy of manipulator mechanisms has been
Accommodation [35], joint compliance [26], active stiffness generally viewed as a problem of resolving the end-effector
desired motion into joint motions with respect to some criteria.
[311, impedance control [9], and hybrid position/force control
[28] are among the various methods that have been proposed. Manipulator redundancy has been aimed at achieving goals
such as the minimization of a quadratic criterion [29], [34], the
Force control has been generally based on kinematic and
avoidance of joint limits [5], [21] the avoidance of obstacles
static considerations. While in motion, however, a manipula-
[4], [6], [20], kinematic singularities [23], or the minimization
tor end-effector is subject to inertial, centrifugal, and Coriolis
of actuator joint forces [8]. The end-effector equations of
motion for a redundant manipulator are established and its
Manuscript receivedAugust 6 , 1985; revised May 29, 1986. This work was behavior with respect to generalized joint forces is described.
supportedinpart by the NationalScienceFoundation and inpart by the The unified approach for motion and active force control is
Systems Development Foundation. then extended to these systems.
The authoris with the Artificial Intelligence Laboratory, Computer Science
Department, Cedar Hall, Stanford University, Stanford, CA 94305. Kinematic singularities is another area that has been
IEEE Log Number 8610323. considered within the framework of joint space control and

.00@1987 IEEE
0882-4967/87/0200-0043$01

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 18:5703 UTC from IE Xplore. Restricon aply.
44 IEEE JOURKAL OF ROBOTICS AND AUTOMATION, VOL. RA-3, NO. FEBRUARY 1987

Fig, 1. Constrainedend-effectorfreedom of motion. (a) Fivedegrees of


freedom. (b) Four degrees of freedom. (c) Three degrees of freedom.

formulated in terms of resolution of the task specifications into lated object forms, with the fixture or constrained object, a
joint motions. Generalized inverses and pseudo-inverses have pair of two rigid bodies linked through a joint. A constrained
been used, and recently an interesting solution based on the motiontask can be described, for instance, by a spherical,
singularity robust inverse has been proposed Inthis planar, cylindrical, prismatic, or revolute joint.
paper, a new approach for dealing with the problem of However, whenviewed from the perspective of end-
kinematic singularities within the operational space framework effector control, two elements of information are required for
is presented. In the neighborhood of a singular configuration a complete description of the task. These are the vectors of
the manipulator is treated as a redundant mechanismwith total force and moment that are to be applied to maintain the
respect to the motion of the end-effector in the subspace of imposed constraints, and the specification of the end-effector
operational space orthogonal to its singular direction. Control motion degrees of freedom and their directions.
of the end-effector for motions along the singular direction is Let f d be a unite vector, in the frame of reference (Ro((3,
based onthe use of the kinematic characteristic of the Jacobian yo, along the direction of the force that is to be applied by
matrix. the end-effector. The positional freedom, if any, of the
constrained end-effector will therefore lie inthe subspace
11. GENERALIZED TASKSPECIFICATION MATRICES orthogonal to f d .
The end-effector motion and contact forces are among the A convenient coordinate frame for the description of tasks
most important components in the planning, description, and involving constrained motion operations is a coordinate frame
control of assembly operations of robot manipulators. The 6if(0,xf, y f , z f ) obtained from Gio by a rotation transforma-
end-effector configuration is represented by a setof m tion described by Sf such that zf is aligned with fd. For tasks
parameters, X I , xz, x,, specifying its positionand where the freedom of motion (displacement) is restricted to a
orientation in some reference frame. In free motion opera- single direction orthogonal to f d , one of the axes Oxf or Oyf
tions, the number of end-effector degrees of freedom mo is will be selected in alignment with that direction, as shown for
defined as the number of independent parameters required the task in Fig. 2.
to specify completely, in a frame of reference (Ro, its position Let us define, in the coordinate @if,the position specifica-
and orientation. A set ofsuch independent configuration tion matrix
parameters forms a system of operational coordinates.
In constrained motion operations, the displacement and
rotations of the end-effector are subjected to a set of geometric
0 0
(1) =f=(T 2
constraints. These constraints restrict the freedom of motion where ox, o;, and a, are binary numbers assigned the value 1
(displacements and rotations) of the end-effector. Clearly, when a free motion is specified along the axes Ox,, Oyf, and
geometric constraints will affect only the freedom of motion of (3zf, respectively, and zero otherwise. A nonzero value of a,
the end-effector, since static forces andmoments at these implies a full freedom of the end-effector position. This case
constraints can still be applied. The number of degrees of of unconstrained end-effector position is integrated here for
freedom of the constrained end-effector is given by the completeness. The coordinate frame ( R f in this case is assumed
difference between mo and the number of independent to be identical to and the matrix Sf is the identity matrix.
equations specifying the geometric constraints, assumed to be The directions of force control are described by the force
holonomic. Examples of five-, four-, and three-degree-of- specification matrix associated with Cf and defined by
freedom constrained end-effectors are shown in Fig. 1.
zf
An interesting description of the characteristics ofend- cf=I-cf (2)
effectors and their constraints uses a mechanical linkage
representation 1241. The end-effector, tool, or manipu- where I designates the 3 3 identity matrix.

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 18:5703 UTC from IE Xplore. Restricon aply.
MOTION KHATIB: CONTROL OF ROBOT
FORCE MANIPULATORS 45

a
The task specification matrices Q and can be constant,
configuration-varying, or time-varying matrices. Nonconstant
generalized task specification matrices correspond to specifi-
cations that involve changes in the direction of the applied
force vector and/or moment vector, e.g., moving the end-
effector while maintaining a normal force to a nonplanar
surface. Cl and have been expressed here with respect to the
frame of reference For control systems implemented for
tasks specified with respect to the end-effector coordinate
frame, these matrices will be specified with respect to that
coordinate frame as well.
EQUATIONS
111. END-EFFECTOR OF MOTION

Joint space dynamic models, which establish the equations


YO
of manipulator joint motions, provide means for the analysis
Fig. 2. One-degree-of-freedom motion.
and control of these motions, and for the description of the
configuration dependency and interactive nature of these
Let us now consider the case where the end-effector task
mechanisms. However, the control of end-effector motion and
involves constrained rotations and applied moments. Let 7 d be
contact forces, or the analysis and characterization of end-
the vector, in the frame of reference of
effector dynamic performance requires the construction of the
moments that are tobe applied by the end-effector, and
model describing the dynamic behavior of this specific part of
z,) be a coordinate frame obtained from
the manipulator system.
by a rotation S, that brings into alignment with the
The end-effector motion is the result of those combined joint
moment vector In a,,the rotation freedom of the end-
effector lies in the subspace spanned by To a task forces that are able to act along or about the axes of
displacement or rotation of the end-effector. These are,
specified in terms of end-effector rotations and applied
indeed, the forces associated with the system of operational
moments inthe coordinate frame a,, we associate the rotation
coordinates selected to describe the position and orientation of
and momentspecification matrices and X,, defined similarly
the end-effector. The construction of the end-effector dynamic
to Cf and gf.
For general tasks that involve end-effector motion(both model is achieved by expressing the relationships between its
positionand orientation) and contact forces (forces and operational positions, velocities accelerations, and the virtual
moments) described in the frame of reference Boywe define operational forces acting on it.
the generalized task specification matrices First, let us consider the case of nonredundant manipula-
tors, where a set of operational coordinates can be selected as a
system of generalized coordinates for the manipulator. The
manipulator configuration is represented by the column matrix
of n joint coordinates, and the end-effector position and
and orientation is described, in a frame of reference by the mo
1 column matrix of independent configuration parame-
ters, i.e., operational coordinates. With the manipulator
nonredundancy assumption we have the equality n mo.
associated with specifications of motion ahd contact forces, Now let us examine the conditions under which a set of
respectively. independent end-effector configuration parameters can be used
Q and fi act on vectors described in the reference frame as a generalized coordinate system for a nonredundant
A position command vector, for instance, initially expressed manipulator. In the reference frame the system of mo
in is transformed by the rotation matrix Sf to the task equations expressing the components of as functions of joint
coordinate frame The motion directions are then selected coordinates, i.e., the geometric model, is given by
in this frame by the application of Cp Finally, the resulting
vector is transformed back in by Sf'. G(g).
The construction of the generalized task specification Let qi and be, respectively, the minimal and maximal
matrices is motivated by the aim of formulating the selection bounds of the ith joint coordinate q i . The manipulator
process in the same coordinate frame (reference frame configuration represented by the point in joint space is
where the manipulator geometric, kinematic, and dynamic confined to the hyperparallelepiped
models are formulated. This allows a more efficient imple-
mentation of the control system for real-time operations.
Control systems using specifications based only
on the
matrices and X, will require costly geometric, kinematic,
and dynamic transformations between the reference frame and Obviously, for arbitrary kinematic linkages, and general joint
the task coordinate frames. boundaries, the set of functions G defined from to the

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 18:5703 UTC from IE Xplore. Restricon aply.
46 IEEE JOURNAL OF ROBOTICS
AUTOMATION,
AND VOL.
NO. RA-3, FEBRUARY 1987

domain 9, of the operational space given by where p ( x , is the vector of end-effector centrifugal and
Coriolis forces given by
pj(x,
X)=X~rIj(x)X, i= 1, mo. (15)
is not one-to-one.
Different configurations of the manipulator links can, in The components of the mo mo matrices rIj(x) are the
fact, be found for a given configuration of the end-effector. Christoffel symbols a i , j k given as a function of the partial
The restriction to a domain where G is one-to-one is therefore derivatives of A(x) with respect to the generalized coordinates
necessary to construct, with the operational coordinates, a x by
system of generalized coordinates for the manipulator mecha-
axjk
nism.
In addition, for some configurations of the manipulator, the
end-effector motion is restricted by the linkage constraints and
The equations ofmotion
1 ah,
aj,jk=-
2 axk
ahik
axj axi 1.
(14) establish the relationships
its freedom of motionlocally decreases. These are the singular
between positions, velocities, and accelerations of the end-
configurations, which can be found by considering the
effector and the generalized operational forces acting on it.
differentiability characteristics of the geometric model G .
The dynamic parameters in these equations are related to the
Singular configurations E are those where the Jacobian
parameters involved in the manipulator joint space dynamic
matrix J ( q ) involved in the variational or kinematicmodel
model. The manipulator equations of motion injoint space are
associated with G ,
given by
J(q)Aq, (8)
q)+g(q)=r (17)
is singular. The end-effector behavior at singular configura-
where b(q, and r represent, respectively, the
tions is treated in Section VIII.
Let 5, be the domain obtained from 9, by excluding the Coriolis and centrifugal, gravity, and generalized forces in
joint space. A ( q ) is the n n joint space kinetic energy
manipulator singular configurations and such that the vector
matrix. The relationship between the kinetic energy matrices
function G of (5) is one-to-one. Let a),designate the domain
A and A(x) corresponding, respectively, to the joint space
S,= (9) and operational space dynamic models can be established 131,
[14] by exploiting the identity between the expressions of the
The independent parameters X I , x2, xmoform a complete quadratic forms of the mechanism kinetic energy with respect
set of configuration parameters for a nonredundant manipula- to the generalized joint. and operational velocities,
tor, in the domain a),of operational space and thus constitute a
system of generalized coordinates for the manipulator system. A(x)=J-*(q)A(q)J-’(q). (18)
The kinetic energy of the holonomic articulated mechanism
is a quadratic form of the generalized operational velocities The relationship between the centrifugal and Coriolis forces
b(q, and p ( x , X) can be established by the expansion of the
1 expression of p(x, X) that results from (1 l),
“(X, X) XTA(x)X (10)
2
p(x, X) h(x)X V T ( x , X ) . (19)
where A(x) designates the mo mo symmetric matrix of the
quadratic form, i.e., the kinetic energy matrix. Using the Using the expression of A(x) in (18), the components of p(x,
Lagrangian formalism, the end-effector equations of motion in (19) can be written as
are given by
A(x)X=J-T(q)A A(q)h(q, 4 ) j - T ( Q ) A ( q ) 4

v m , X) J- =(q>l(q, j-* ( S ) A (20)


where the Lagrangian L ( x , X) is where

L X) T ( x , X) U ( x ) (12) h(q, 4 ) j ( q M (21)


and U ( x ) represents the potential energy due to gravity. F is and
the operational force vector. Let p ( x ) be the vector of gravity
1
forces li(Q, q=Aqi(q)q, i=l, n. (22)
2
p ( x ) VU(x). (13)
The subscript q i indicates the partial derivative with respect to
The end-effector equations of motion in operational space can the ith joint coordinate. Observing from the definition of b(q,
be written 131, [14] in the form that

A ( xX) f) + p ( ~x ,) = F b(q1 4>=A(s)4-4s,4) (23)

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 18:5703 UTC from IE Xplore. Restricon aply.
KHATIB: MOTION AND FORCE CONTROL OF ROBOT MANIPULATORS 47

yields Nonlinear dynamic decoupling in operational space is


obtained by the selection of the following control structure,
4 J- =(q)b(q, Nq)h(q, (24)
F = Fm Fccg (29)
The relationship between the expressions of gravity forces can
be obtained using the identity between the functions expressing with
the gravity potential energy in the two systems of generalized
coordinates and the relationships between the partial deriva- Fm
tives with respect to these coordinates. Using the definition of
the Jacobian matrix (8) yields

and j represent the estimates of where


In the foregoing relations, the components involved inthe end- A(x), and p FZ is the command vector of the
effector equations of motion (14), i.e., A, p , are expressed decoupled end-effector. With a perfect nonlinear dynamic
in terms of joint coordinates. This resolves the ambiguity in decoupling, the end-effector becomes equivalent to a. single
defining the configuration of the manipulator corresponding to unitmass moving in the mo-dimensional space. To
a configuration of the end-effector in the domain of -(7). simplify the notations, the symbol will be dropped in the
With these expressions, the restriction to the domain 6,, following development.
where is one-to-one, then becomes unnecessary. Indeed, the At the level of the decoupled end-effector, I$, various
domain of definition of the end-effector dynamic model of a control structures can be selected. For tasks where the desired
nonredundant manipulator can be extended to the domain a), motionof the end-effector is specified, a linear dynamic
defined by behavior can be obtained by selecting

a),) (26) (31)

where is the domain resulting from a ),of (6) by excluding where and are the desired position, velocity, and
the kinematic singular configurations. acceleration, respectively, of the end-effector. is the mo
Finally, let us establish the relationship between generalized mo identity matrix. kp and k, are the position and velocitygain
forces, i.e., F and r. Using (18), (24), and (25) the end- matrices.
effector equations of motion (14) can be rewritten as An interesting approach for tasks that involve large motion
to a goal position, where a particular path is not required, is
J - T(q)rA +g(q)l= F. (27) based on the selection of the decoupled end-effector command
Substituting (17) yields vector F; as

r J T ( q )F (28) FZ k,(i- (32)


which represents the fundamental relationship between opera- where
tionaland joint forces consistent with the end-effector and
manipulator dynamic equations. This relationship is the basis
for the actual control of manipulators in operational space.
MOTION
IV. END-EFFECTOR CONTROL
(33)
The control of a manipulator in operational space is based
on the selection of the generalized operational forces F as a
command vector. These forces are produced by submitting the This allows a straight line motion of the end-effector at a given
manipulator to thecorresponding joint forces I' obtained from speed V,=. The velocity vector is in fact controlled to be
(28). pointed toward the goal position while its magnitude is
As with joint space control systems, the control in opera- limited to V,,,. The end-effector will then travel at V,,, in a
tional space can be developed using a variety of control straight line, except during the acceleration and deceleration
techniques. In operational space control systems, however, segments. This command vector is particularly useful when
errors, performance, dynamics, simplifications, characteriza- used in conjunction with the gradient of an artificial potential
tions, and controlled variables are directly related to manipula- field for. collision avoidance 151.
tor tasks. Using the relationship between generalized forces given in
One of the most effective techniques for dealing with these (28), the joint forces corresponding to the operational com-
highly nonlinear and strongly coupled systems is the nonlin- mand vector F, in (29) and (30), for the end-effector dynamic
eardynamicdecouplingapproach [36],[37], which fully decoupling and control, can be written as
exploits the knowledge of the dynamic model structure and
parameters. Within this framework of control and at the level
r =J=(q)A(q)F: (34)
of the uncoupled system linear, nonlinear, robust [32], and where is the vector of joint forces under the mapping
adaptive [3] control structures can be implemented. into joint space of the end-effector Coriolis and centrifugal

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 18:5703 UTC from IE Xplore. Restricon aply.
48 JOURNAL
IEEE OF ROBOTICS AND AUTOMATION, VOL. RA-3, NO. I , FEBRUARY 1987

force vector To simplify the notation, A has also been This approach has also been proposed [lo] for real-time
used here to designate the kinetic energy matrix when dynamic control of manipulators in joint space.
expressed as a function of the joint coordinate vector
4 ) is distinct from the vector of centrifugal and Coriolis forces V. CONSTRAINED MOTION OPERATIONS
b(q, in(17)that arises whenviewingthe manipulator The matrix s2 defined earlier specifies, with respect to the
motion in joint space. These vectors are related by frame of reference 030, the directions of motion (displacement
and rotations) of the end-effector. Forces and moments are to
4) 4 ) J T ( q ) A ( q ) h ( q ,4 ) . (35) be applied in or about directions that are orthogonal to these
A useful form of for real-time control and dynamic motion directions. These are specified by the matrix 8.
analysis can be obtained by a separation of its dependency on An important issue related to the specification of axes of
position and velocity. rotation and applied moments is concerned with the compati-
The joint space centrifugal and Coriolis force vector b(q, bility between these specifications and the type of representa-
of (17) can, in fact, be developed in the form tion used for the description of the end-effector orientation. In
fact, the specification of axes of rotations and applied moments
b(q7 4 ) =B(q)[441+ c(q)[q21 (36) in the matrices and 2, are only compatible with descriptions
of the orientation using instantaneous angular rotations.
where B ( q ) and are, respectively, the n n(n 1)/2 However, instantaneous angular rotations are not quantities
and n n matrices of the joint space Coriolis and centrifugal that can be used as a set of configuration parameters for the
forces associated with b(q, 4 ) . [qq] and [q2]are the symbolic orientation. Representations of the end-effector orientation
notations for the n(n 1)/2 1 and n 1 column matrices such as Euler angles, direction cosines, or Euler parameters,

[441=[4142 4143 4n-1@nIT


are indeed incompatible with specificationsprovided by and
E,.
[42] [q: Qi]T. (37) Instantaneous angular rotations havebeenused for the
description of orientation error of the end-effector. An angular
With and [g2],the vector h(q, can be developed in the rotation error vector 64 that corresponds to the error between
form the actual orientation of the end-effector and its desired
orientation can be formed from the orientation description
h(q7 4 ) = ~ l ( 4 ) [ 9 Q 1 + ~ 2 ( ~ ) c 4 2 1 (3 8)
given by the selected representation 131, [22].
where the matrices .Hl(q)and H2(q)have, respectively, the The time derivatives of the parameters corresponding to a
dimensions n n( 1)/2 and n n. Finally, the vector representation of the orientation are related simply to the
g(q, can be written as angular velocity vector. With linear and angular velocities is
associated the matrix Jo(q), termed the basic Jacobian,
4 ) =B(q)[(i@l+
aq)[421 (39) defined independently of the particular set of parameters used
where B ( q ) and c(q)are the n n(n 1)/2 and n n to describe the end-effector configuration
matrices of the joint forces under the mapping into joint space
of the end-effector Coriolis and centrifugal forces. These
matrices are
The Jacobian matrix J ( q ) associated with a given representa-
B(q)=B(q)-JT(g)A(q)H1(q) tion of the end-effector orientation x, can then beexpressed in
the form [13]
JT(q>A(q)H2(q) (40)
Withthe relation (39), the dynamic decoupling of the end- J ( q ) ExrJo(4) (43)
effector can be obtained using the configuration dependent where the matrix Ex, is simply given as a function of x,.
dynamic coefficients A ( q ) , B ( q ) , and The joint c(q), For end-effector motions specified in terms of Cartesian
force control vector (34) becomes coordinates and instantaneous angular rotations, the dynamic
decouplingandmotion control of the end-effector can be
~ = J ~ ( ~ ) A ( ~ ) F I T , +e B( m( ~2 )i +[ ~m~ .I (41)
achieved 131 by
By isolating these coefficients, end-effector dynamic decou-
pling and control can be achieved in a two-level control system r J,T(s>Ao(x)F: 4 )+ g ( 4 ) (44)
architecture The real-time computation of these coeffi- where and are defined similarly to A ( q ) and
cients can then be pacedby the rate of configuration changes, 4 ) with J ( q ) being replaced by Jo(q).
which is much lower than that of the mechanism dynamics. Using the relationship (43) similar control structures can be
This leads to the following architecture for the control system designed to achieve dynamic decoupling and motion control
a low rate dynamic parameter evaluation level: updat- with respect to descriptions using other representations for the
ing the end-effector dynamic parameters; orientation of the end-effector
a high rate servo control level: computing the command The unified operational command vector for end-effector
vector (41) using the updated dynamic coefficients. dynamic decoupling, motion, and active force control can be

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 18:5703 UTC from IE Xplore. Restricon aply.
KHATIB: MOTION AND FORCE CONTROL OF ROBOT MANIPULATORS 49

Fig. 3. Operationalspacecontrol system architecture.

written as manipulator, and the dynamic behavior of the entire redundant


system cannot be represented by a dynamic model in coordi-
F = Fm Fa+ Fccg (45) natesonly of the end-effector configuration. The dynamic
where Fm, Fa, and Fccgare the operational command vectors of behavior of the end-effector itself, nevertheless, can still be
motion, active force control, and centrifugal, Coriolis, and described, and its equations of motionin operational space can
gravity forces given by still be established.
The end-effector is affected by forces acting along or about
F m Ao(Q)Qp: the axes of its freedom of motion. These are the operational
forces associated with the operational coordinates selected to
Fa=fiFz+Ao(q)fiFz describe its position and orientation. Let us consider the end-
Fccg 4) (46) effector dynamic response to the application, on the end-
effector,oran operational force vector F. In this case of
where F: represents the vector of end-effector velocity redundant manipulator systems, the joint forces that can be
damping that acts in the direction of force control. The joint used to produce a given operational force vector are not
force vector corresponding to F in (45) is unique. The joint force vector
r=J,T(q)[Ao(q)(QF: OF:) fiFzI+ 4) I'=JT(q)F (28)
(47) represents, in fact, one of these solution.
The control system architecture is shown in Fig. 3, where kf The application of the joint forces (28) to the manipulator
represents the force error gain and k,f denotes the velocity (17), and the use of the relation
gain in F,*. An effective strategy for the control of the end- a=J(q)q+h(q, (49)
effector during the transition from free to constrained motions
is based ona pure dissipation of the energy at the impact. The allow us to establish [13] the equations of motion of the end-
operational command vector Fa during the impact transition effector
control stage is
4 ) +PA41 = F
Fa=Ao(q)8F:. (48)
where
The duration of the impact transition control is a function of
the impact velocity and the limitations on damping gains and h,(q)=[J(q)A-'(q)JT(q)l-'
actuator torques (this duration is typically on the order of tens
of milliseconds). Force rate feedback has also been used inF;.
4 ) = J T ( q ) b ( q 94 ) - & ( 4 ) h ( q , 4)
A more detailed description of the components involved in this PA41 .fT(4k(q>
control system, real-time implementation issues, and experi-
mental results can be found in [19]. and

VI. REDUNDANT
MANIPULATORS ~i(4)=A-'(s)JT(4)A,((I).
The configuration of a redundant manipulator cannot be J ( q ) actually a generalized inverse of the Jacobian matrix
specified by a set of parameters that only describes the end- corresponding to the solution that minimizes the manipulator's
effector position and orientation. An independent set of end- instantaneous kinetic energy.
effector configuration parameters, therefore, does not consti- Equation (50) describes the dynamic behavior of the end-
tute a generalized coordinate system for a redundant effector when the manipulator is submitted to a generalized

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 18:5703 UTC from IE Xplore. Restricon aply.
50 IEEE JOURNAL OF ROBOTICS AND AUTOMATION, VOL.
NO. RA-3, FEBRUARY 1987

joint force vector of the form (28). The m m matrix Stability Analysis
can be interpreted as a pseudo-kinetic energy matrix corres- In the command vector and with the assumption of a
pending to the end-effector motion in Operational 'pace* "perfect" compensation (or noncompensation) ofthe centrifu-
0)represents the and forces acting On the gal and Coriolis forces, the manipulator can be consideredas a
end-effector and p r ( q )the gravity force vector.
conservative system subjected to the dissipative forces due to
The effect on the end-effector of the application of arbitrary the velocity damping term kui) in F.+. These forces are
joint forces, can be determined by (50) which can berewritten

Substituting 7 ) yields D(q) k u J T ( q ) ~ r ( q ) J ( q( 6) .0 )


~=~=(q)r. Lyapunov stability analysis leads to the condition
This relationship determines how the joint space dynamic qTD(q)qsO (61)
forces are reflected at the level of the end-effector.
Lemma: The unconstrained end-effector (50) is subjected to whichis satisfied, since D ( q ) is an n n negative
the operational force F if and only if the manipulator (17) is semidefinite matrix of rank m. However, the redundant
submitted to the generalized joint force vector mechanism can still describe movements that are solutions of
the equation
r = J T ( q ) ~ + [ z , - J T ( q ) J T ( q ) ] r o (55)
qTD(q)q=o. (62)
where I,, is the n n identity matrix, J ( q ) is the matrix given
in (52) and r0is an arbitrary joint force vector. An example of such behavior is shown in Fig. 4(a). The end-
When the applied forces r are of the form (55), it is effector of a simulated three-degree-of-freedom planar manip-
straightforward from to verify that the only forces acting ulator is controlled under (58). The end-effector goal position
on the end-effector are the operational forces F produced by coincides with its current position, while the three joints are
the first term in the expression of r. Joint forces of the form assumed to have initially nonzero velocities (0.5 rad/s has
[I,, J T ( q ) J T ( q ) ] rcorrespond
0 in fact to a null operational been used).
force vector. Asymptotic stabilization of the system can be achieved by
The uniqueness of (55) is essentially linked to the use of a the addition of dissipative joint forces [ 1 3 ] . These forces can
generalized inverse that is consistent with the dynamic be selected to act in the null space of the Jacobian matrix 161.
equations of the manipulator and end-effector. The form of the This precludes any effect of the additional forces on the end-
decomposition (55) itself is general. A joint force vector r can effector and maintains its dynamic decoupling. Using (55)
always be expressed in the form of (55). these additional stabilizing joint forces are of the form
Let P ( q )be a generalized inverse of J ( q ) , and let us submit
rn,=[Z,-JT(q)JT(q)Ir,.
(63)
the manipulator to the joint force vector
By selecting

rs=
(64)
-ku,A(q)4,
If, for any ro,the end-effector is only subjected to F, (56) the vector r,, becomes
yields
rns=rs+JT(q)Ar(q)Frs
(65)
. I ( q ) A - ' ( q ) = [ J ( q ) A - ' ( q ) J T ( q ) 1 P T ( q(57)
)
with
which implies the equivalence of P ( q ) and J ( q ) . (66)
Fr, kuqi.
VII. CONTROL
OF REDUNDANT
MANIPULATORS Finally, the joint force command vector can be written as
As in the case of nonredundant manipulators, the dynamic
decoupling and control of the end-effector can be achieved by r=J'(q)ar(q)(F.+.+Fr,)+r,+b;(q, q ) + g ( q ) -( 6 7 )
selecting an operational command vector of the form of (29),
(30): The corresponding joint forces are Under this form, the evaluation of the generalized inverse of
the Jacobian matrix is avoided. The matrix D ( q ) correspond-
ing to the new expression for the dissipativejoint forces rdis
in
the command vector (67) becomes
where is defined similarly to 4). D ( q ) = - [ ( k , - k u , ) J T ( q ) A , ( q ) J ( q ) + k u , A ( q ) l (68)
.
The manipulator joint motions produced by this command
vector are those that minimize the instantaneous kinetic energy Now, the matrix D ( q ) is negative definite and the system is
of the mechanism. asymptotically stable. Fig. 4(b) shows the effects of this
Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 18:5703 UTC from IE Xplore. Restricon aply.
KHATIB: MOTION AND FORCE CONTROL OF ROBOT MANIPULATORS 51

(a) (b)
Fig. 4. Stabilization of redundant manipulator.

stabilization on the previous example of a simulated three- hood of a given type of singularity 9, can be defined as
degree-of-freedom manipulator.
Constrained Motion Control
as=~ ~ 1 1 s ( q ) l ~ ~ o l (70)

The extension to redundant manipulators of the results where is positive.


obtained in the case of nonredundancy is straightforward. The The basic concept in our approach to the problem of
generalized joint forces command vector becomes kinematic singularities canbe formulated as follows. In the
neighborhood 9, of a singular configuration the manipula-
r J~(q)[ii,,(q)(w; i j ~+F,J+
; ij~;] tor is treated as a mechanism that is redundant with respect to
the motion of the end-effector in the subspace of operational
r, gr0(q, (69)
space orthogonal to the singular direction. For end-effector
where and b;o(q, are defined with respect to the motioninthat subspace, the manipulator is controlled as a
basic Jacobian matrix Jo(q). redundant mechanism. Joint forces selected.from the associ-
ated null space are used for the control of the end-effector
CONFIGURATIONS
VII. SINGULAR motion along the singular direction. When moving out of the
A singular configuration is a configuration q at which singularity, this is achieved by controlling the rate of change
some column vectors of the Jacobian matrix become linearly of according to the value of the desired velocity for this
dependent. The mobility of the end-effector can be defined as motion at the configuration when Selecting the
the rank of this matrix In the case of nonredundant sign of the desired rate of change of allows the control of
manipulators considered here, the end-effector at a singular the manipulator posture among the two configurations that it
configuration loses the ability to move along or rotate about can generally take whenmovingoutof a singularity. A
some direction of the Cartesian space; its mobilitylocally position error term on is used in the control vector for
decreases. Singularity and mobility can be characterized, in tasks that involve a motion toward goal positions located at or
this case, by the determinant of the Jacobian matrix. in the neighborhood of the singular configuration.
Singularities can be further specified by the posture.of the Using polar or singular value decomposition, this approach
mechanism at which they occur. Different types of singulari- can be easily extended to redundant manipulator systems. The
ties can be observed for a given mechanical linkage. These can extension to configurations where more than one singularity is
be directly identified from the expression of the determinant of involved can be also simply achieved. An example of a
the Jacobian matrix. The expression of this determinant can, in simulated two-degree-of-freedom manipulator is shown in
fact, be developed into a product of terms, each of which Fig. 5(a). The manipulator has been controlled to move into
corresponds to a type of singularity related to the kinematic andoutof the singular configuration while displaying two
configuration of the mechanism, e.g., alignment of two links different postures. The time-response of the motionin the
or alignment of two joint axes. singular direction x(t) is shown in Fig. 5@).
To each singular configuration there corresponds a singular
“direction.” It is in this direction that the end-effector IX. SUMMARY
AND DISCUSSION

presents infinite inertial mass for displacements or infinite A methodology for the description of end-effector con-
inertia for rotations. Its movements remain free in the strained motion tasks based on the construction of generalized
subspace orthogonal to this direction. This behavior extends, task specification matrices has been proposed. For such tasks
in reality, to a neighborhood of the singular configuration. The where both motion and active force control are involved, a
extentof this neighborhood canbe characterized by the unified approach for end-effector dynamic control within the
particular expression in the determinant of the Jacobian operational space framework has been presented. The use of
matrix that vanishes at this specific singularity. The neighbor- the generalized task specification matrix has provided a more

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 18:5703 UTC from IE Xplore. Restricon aply.
52 IEEE JOURNAL OF ROBOTICS AND AUTOMATION, VOL. RA-3, NO. I , FEBRUARY 1987

Fig. 5. Control at singular configuration.

efficient control structure for real-time implementations, alternative to resolving end-effector motions intojoint motions
further enhanced by a two-level control architecture. generally used in joint space based control systems.
Results of the implementation of this approach have shown
the operational space formulation to be an effective means of ACKNOWLEDGMENTS
achieving high dynamic performance in real-time motion The author would like to thank Brian Armstrong, Harlyn
control and active force control of robot manipulator systems. Baker, Joel Burdick, Bradley Chen, John Craig, Ron Fearing,
This approach has beenimplementedin an experimental Jean Ponce, and Shashank Shekhar for the discussions,
manipulator programming system COSMOS (Control in Oper- comments, and help during the preparation of the manuscript.
ational Space of a Manipulator-with-ObstaclesSystem): Using
a PUMA 560 and wrist and finger sensing, basic assembly REFERENCES
operations have been performed. These include contact, slide, [l] B. Armstrong, Khatib, and J. Burdick, “The explicitdynamic
model and inertial parameters of the PUMA 560 arm,” in Proc. I986
insertion, and compliance operations [17]. With the recent ZEEE Znt. Conf. Robotics and Automation, pp. 510-518.
implementation of COSMOS in the multiprocessor computer [2]B. Chen, R. Fearing, B. Armstrong, and J. Burdick, “NYMPH: A
systemNYMPH [2], where four National Semiconductor multiprocessor for manipulation applications,” in Proc. I986 ZEEE
Znt. Conf. Robotics and Automation, pp. 1731-1736.
32016 microprocessors have been used, a low-level servo rate [3] J. J. Craig, P. Hsu, S . S . Sastry, “Adaptive controlofmechanical
of 200 Hz and a high-level dynamics rate of 100 Hz have been manipulators,” in Proc. 1986 ZEEEZnt. Conf. Robotics and
achieved. Automation, pp.190-195.
[4] B. Espiau and R. Boulic, “Collision avoidance for redundantrobots
The impact transition control strategy was effective in the withproximity sensors,” in Preprints 3rdZnt. Symp. Robotics
elimination of bounces at contact with a highly stiff surface. Research, 1985, pp. 94-102.
The end-effector normal velocities at impact were up to 4.0 in/ [5] A. Fournier, “Gtntrationde mouvements en robotique-application
des inverses gtntralisbeset des pseudo inverses,” Thtse d’etat,
s. Experiments with square wave force input have also been Mention Science, Univ. des Sciences et TechniquesduLanguedoc,
conducted, and responses with rise times of less than 0.02 s Montpellier, France, Apr. 1980.
[6] H. Hanafusa, T. Yoshikawa, and Y. Nakamura, “Analysis and control
andsteady force errors of less than 12 percent have been of articulated robot arms with redundancy,” in Proc. 8th ZFAC World
observed. This performance hasbeen obtained despite the Congress, vol. XIV, 1981, pp.38-83.
limitations in controlling the manipulator joint torques [27]. J. R. Hewit and J . Padovan, “Decoupled feedback control of a robot
and manipulator arms,” in Proc. 3rd CZSM-ZFToMMSymp. Theory
Accurate identification of the PUMA 560 dynamic parameters and Practice of Robots and Manipulators. NewYork: Elsevier,
[l] has contributed to a nearly perfect dynamic decoupling of 1979, pp. 251-266.
the end-effector. [SI J. M. Hollerbach and K. C. Suh, “Redundancy resolution
of
manipulators through torque optimization,” in Proc. I985 Znt. Conf.
For redundant manipulator systems, the end-effector equa- Robotics and Automation, pp. 1016-1021.
tions of motion have been established, and an operational [9] N. Hogan, “Impedance control of a robotic manipulator,” presented at
space control system for end-effector dynamic decoupling and the Winter Annual Meeting of the ASME, Washington, DC, 1981.
[lo]A. Izaguirre, and R. P. Paul,“Computationoftheinertialand
control has been designed. The expression of joint forces of gravitational coefficients of the dynamic equations for a robot manipu-
the null space of the Jacobian matrix consistent with the end- lator with a load,” in Proc. I985 Znt. Conf. Robotics and
effector dynamic behavior has been identified and used for the Automation, pp.1024-1032.
[ll] Khatib, M. Llibre, and R. Mampey, “Fonction dtcision-commande
asymptotic stabilization of the redundant mechanism. The d’un robot manipulateur,” Rapport Scientifique 2/71-56, DERA-CERT,
resulting control system avoids the explicit evaluation of any Toulouse, France, July 1978.
generalized inverse or pseudo-inverse of the Jacobian matrix. [12] Khatib and J. F.Le Maitre, “Dynamic control of manipulators
operating in a complex environment,” in Proc. 3rd CZSM-ZFToMM
Joint constraints, collision avoidance 121, 151, and control of Symp, Theory andPractice of Robots and Manipulators. New
manipulators’ postures can be naturally integrated in this York: Elsevier, 1979,pp.267-282.
framework of operational space control systems. Also, a new [13] Khatib, “Commande dynamique dans I’espace optrationnel des
robots macipulateurs en prtsence d’obstacles,” these de docteur-
systematic solution to the problem of kinematic singularities ingtnieur, Ecole National Suptrieure de 1’ACronautique et de I’Espace
has been presented. This solution constitutes an effective (ENSAE). Toulouse, France, 1980..

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 18:5703 UTC from IE Xplore. Restricon aply.
KHPrTIB: MOTION AND FORCE CONTROL OF ROBOT MANIPULATORS 53

141 “Dynamiccontrol of manipulators in operationalspace,” in [30]M. Renaud and J.Zabala-Iturralde,“Robotmanipulatorcontrol,” in


Proc. 6th CZSM-ZFToMM Congress on Theory of Machines and Proc. 9th Int. Symp. Industrial Robots, 1979, pp. 463-475.
Mechanisms. New York: Wiley,1983, pp. 1128-1131. [31] J. K. Salisbury, “Active stiffness control of a manipulator in Cartesian
“Real-timeobstacleavoidancefor manipulators and mobile coordinates,” presented at the 19th IEEE Conf. Decision and Control,
robots,” in Proc. ZEEE Int. Conf. Robotics and Automation, pp. Albuquerque, NM, Dec. 1980.
500-505. [32] J. J. Slotine and 0. Khatib, “Robust control in operational space for
161 “Operationalspace formulation in the analysis,design, and goal positioned manipulator tasks,” presented at the 1986 American
control of manipulators,” in Preprints 3rd Znt. Symp. Robotics Control Conf., Seattle, WA, June 1986.
Research, 1985, pp. 103-110. [33] K.Takase,“Task-orientedvariablecontrol of manipulator and its
171 Khatib, J. Burdick, and B. Armstrong, “Robotics in three acts- softwareservoingsystem,” in Proc. ZFACZnt. Symp., 1977, pp.
Part (film), Artificial Intell. Lab., Stanford Univ., Stanford, CA, 139-145.
1985. [34] D. E. Whitney, “Resolved motion ratecontrol of manipulators and
0.Khatib and J. Burdick, “Dynamic optimization in manipulator human prostheses,” ZEEE Trans. Man-Machine Syst., vol. MMS-2,
design: Theoperatioqalspaceformulation,” presented at the 1985 pp. 47-53, June 1969.
ASME Winter Annual Meeting, Miami, FL. [35] D. E. Whitney, “Force feedbackcontrol of manipulator fine motions,”
“Motion and force control of robot manipulators,”, in Proc. ASME J. Dynamic Syst., Meas., Contr., pp. 91-97, June 1977.
1986 ZEEE Znt. Conf. Robotics and Automation, pp. 1381-1386; [36] J. Zabala-Iturralde, “Commande des robots manipulateurs B partir de
M. Kircanski and M. Vukobratovic, “Trajectory planning for redun- la modelisation de leur dynamique,” these de 3””” cycle, Univ. Paul
dant manipulators in presence of obstacles,” in Preprints 5th CZSM- Sabatier, Toulouse, France, July 1978.
ZFToMM Symp. Theory andPractice of Robots and Manipulators, [37] E. Freund,“Thestructure of decoupled nonlinear systems,” Znt. J.
1984, pp. 43-58. Contr., vol. 21, no. 3, pp. 443-450, 1975.
A. Liegois, “Automation supervisory control of the configuration and
behavior of multibody mechanisms,” ZEEE Trans. Syst., Man,
Cybern., vol. SMC-7, Dec. 1977.
J. Y. S. Luh,M.W.Walker, and P. Paul, “Resolved acceleration
control of mechanical manipulators,” ZEEE Trans. Automat. Contr.,
pp. 468-474, June 1980.
J. Y. S. Luh and Y. L. Gu, “Industrial robots with seven joints,” in
Proc. 1985ZEEEZnt. Conf. Robotics and Automation, pp. 1010-
1015.
M. T. Mason, “Compliance and force control for computer controlled
manipulators,” TEEE Trans. Syst., Man, Cybern., vol. SMC-11, pp.
418-432, 1981.
Y. Nakamura, “Kinematical studies on the trajectory control of robot
manipulators,” Ph.D. dissertation, Kyoto, Japan, June 1985.
R.P. Paul and B. Shimano, “Compliance and control,” in Proc. Joint
Automatic Control Conf., pp. 694-699, July 1976.
L. Pfeffer, Khatib, and J. Hake, “Joint torque sensory feedback in
the control of a PUMA manipulator,” presented at the 1986 American
Control Conf., Seattle, WA, June 1986.
M. Raibert and J. Craig, “Hybrid position/force control of manipula-
tors,’’ ASME J. Dynamic Syst., Meas., Contr., June 1981.
M. Renaud, “Contribution B 1’6tude dela modklisation et de la
commmande des systemes mkcaniques articules,” these de docteur
ingenieur, Univ. Paul Sabatier, Toulouse, France, Dec. 1975.

Authorized licensd use limted to: IE Xplore. Downlade on May 10,2 at 18:5703 UTC from IE Xplore. Restricon aply.

Vous aimerez peut-être aussi