 |
Description  |
|
|
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to robots controlled by a computer system in
order to urge a tool, integral with a robot's arm, to follow a memorized
path.
2. Discussion of the Related Art
Many robots are capable of moving a tool, such as milling cutters, torch
welders, etc., within a prememorized precise path. In order to move a tool
according to n degrees of freedom, a robot must have at least n axes. A
degree of freedom corresponds to a translation or rotation movement in
accordance with an axis of a three-dimensional referential system. With 6
degrees of freedom, all the translation and rotation combinations of a
tool are possible (the tool can be positioned anywhere, according to any
orientation within a determined range of action of the robot). The
relation between the articulation variables (rotation angles of the
robot's axes) and the position and orientation of the tool is given by:
P=f(A)
where P is a vector whose components represent the position coordinates
x.sub.i (i=1 . . . 3) and orientation coordinates r.sub.i (i=1 . . . 3) of
the tool, A is a vector whose components represent the articulation
variables a.sub.j (j=1 . . . 6), and f is a vector function.
A slight variation dA of the articulation variables about vector A is then
associated with a slight variation dP of the position vector by the
relation
dP=[J(A)]dA,
where [J(A)] is a Jacobian matrix whose coefficients are functions of the
components of vector A.
Considering the accuracy of trajectories obtained with robots, it would
seem advantageous to use them in the surgical field. However, when the
present robots are operating, they exhibit hazards that render them
unsuitable for surgical treatments. Indeed, in case of failure of the
control circuit of the robot's motor, or simply because of power supply
trouble, such as a brief current interruption, the robot could make an
uncontrolled movement of large magnitude and high power that would
endanger a patient withstanding a surgical act determined by the robot.
Nowadays, robots are used in the surgical field only for prepositioning and
are blocked during the surgical act. For example, a neuro-surgical act
consists in inserting a needle into the brain at a very precise spot. To
achieve this purpose, one uses a robot that positions a cylinder guiding a
needle in a very precise way with respect to the patient's head, using a
system allowing to detect the position of the head. Once the guiding
cylinder is positioned (far enough from the patient so that he cannot be
reached by an uncontrolled movement of the robot), the robot's axes are
blocked and its power supply is turned off. Then, the surgeon inserts the
needle in the desired way into the guiding cylinder. Thus, in the surgical
field, the use of robots is presently limited to the detection of a
precise initial position.
SUMMARY OF THE INVENTION
An object of the invention is to provide a new design of robots capable of
urging a tool to follow precise paths without any risk of uncontrolled
movements.
Another object of the invention is to provide a robot whose movements are
directly controlled by an operator without resorting to forces other than
the one provided by the operator.
Another object of the invention is to provide a robot capable of passively
guiding the operator's movements.
These objects are achieved with a robot operable to guide movements and
whose power is provided by an operator who handles a tool integral with
time robot. Like a conventional robot, the guiding robot has several
degrees of freedom (or axes), respectively associated with movement
sensors that are connected to a computer system in which is stored a path
to be followed. When the operator moves the tool beyond the stored path,
suitable blocking devices, controlled by the computer system and replacing
the robot's motors, prevent the tool from continuing to move beyond its
predetermined path. Thus, at any time, the operator can move the tool only
according to the stored path. The blocking devices are passive, that is,
they are such that they can only resist to a movement but cannot generate
a movement, which prevents the occurrence of any uncontrolled movement.
The invention more particularly relates to a robot having several degrees
of freedom, each associated with a movement sensor that is connected to a
computer system calculating, from information provided by each sensor, the
position of a tool integral with an arm of the robot, the computer system
further memorizing a path. According to the invention, the tool is handled
by an operator and the robot includes a device for determining the
direction of the movement associated with each degree of freedom. The
device authorizes movements according to the associated degree of freedom
in a predetermined direction, the opposite direction, both directions, or
no direction at all, as a function of signals provided by the computer
system.
According to an embodiment of the invention, the device for determining the
direction allows movements according to the associated degree of freedom
only in a direction suitable to urge the tool to penetrate into an
authorized area, that may be limited to a single point, if the tool moves
beyond the area boundary.
According to an embodiment of the invention, the device for determining the
direction allows movements according to the associated degree of freedom
only in a direction suitable to bring the tool closer to the stored path
according to an oblique direction.
According to an embodiment of the invention, the device for determining the
direction includes two free wheels having an opposite rotation direction
and each including a means controlled by the computer system coupled to an
axis constituting the associated degree of freedom.
According to an embodiment of the invention, each coupling means includes a
clutch controlled by a solenoid.
According to an embodiment of the invention, the robot includes, for each
degree of freedom, a speed limiting means, controlled by the computer
system.
The invention also provides a method for controlling a robot having several
degrees of freedom and provided with movement sensors, operable to urge a
tool, integral with a robot's arm, to follow a path memorized in a
computer system, and includes the following steps: manually moving the
tool; calculating the current position of the tool and, if required,
determining as a function of this position a preferential direction to be
followed by the tool; and authorizing a movement according to each degree
of freedom of the robot only in a direction allowing the operator to move
the tool according to the preferential direction.
According to an embodiment of the invention, the movement direction for
each degree of freedom is determined by the sign of a component according
to this degree of freedom of a vector of the space of movements according
to the degrees of freedom, corresponding to a vector indicating, in the
space of positions and orientations of the tool, the preferential
direction.
According to an embodiment of the invention, the preferential direction is
perpendicular to the boundary of an authorized area, that can be limited
to a single point, if the tool moves beyond the boundary.
According to an embodiment of the invention, the preferential direction is
comprised between the perpendicular and the tangent to the stored path.
The foregoing and other objects, features, aspects and advantages of the
invention will become apparent from the following detailed description of
the present invention which should be read in conjunction with the
accompanying drawings.
BRIEF DESCRIPTION OF DRAWINGS
FIG. 1 shows an embodiment of a robot's axis according to the invention;
and
FIGS. 2A and 2B show two positions of a two-axis robot according to the
invention, useful for illustrating its operation.
FIG. 3 shows a graph which is useful for explaining an aspect of the
invention.
DETAILED DESCRIPTION
FIG. 1 shows an axis A of a robot according to the invention. Axis A is for
example integral with an arm 10 and connected to the output of a first
reducing gear 12 integral with a framework B forming a portion, for
example, of a second arm of the robot. The input axis A' of the reducing
gear is connected to a brake 14 controlled by a signal F, to a position
sensor such as a coding wheel 16 associated with an optical sensor 17
providing a rotation angle signal a, and to the output of a second
reducing gear 20. A third reducing gear can be provided between the brake
and the position sensor.
The elements described up to now are conventionally found in robot's
articulations. Conventionally, the input axis A" of reducing gear 20 is
connected to a motor.
According to an embodiment of the invention, axis A" is connected to two
disengageable free wheels 22, 23. The free wheel 22 allows the rotation of
axis A, through reducing gears 12, 20, in only one direction S- when a
control signal C- is active. The free wheel 23 allows the rotation of the
axis only in the opposite direction S+ when a control signal C+ is active.
When the control signal C associated with a free wheel is OFF, the free
wheel allows axis A to rotate in both directions.
With this configuration, axis A provides four functions:
freely rotating in directions S+ and S- when signals C+ and C- are OFF;
rotating only in direction S+ when signal C+ is ON and signal C- OFF;
rotating only in the opposite direction S- when signal C- is ON and signal
C+ is OFF; and
remaining blocked if signals C+ and C- are ON.
Brake 14 provides either one of two different functions, that is, limiting
the rotation speed of axis A to a value predetermined by signal F, or
applying a braking force predetermined by signal F. Brake 14 is normally
blocked so as to block the robot in case of power failure. Similarly, for
double safety, the disengageable free wheels 22, 23 are normally engaged
(signals C+ and C- are ON when the power is turned off). In order to be
able to move the robot in case of power failure, the brakes and free
wheels are designed so as to be manually unblockable.
As an example of a normally-engaged disengageable free wheel is a free
wheel coupled to a clutch controlled by a solenoid during the
disengagement phase.
A computer system 25 stores data such as points of a path or of a surface.
As a function of signals a provided by the rotation sensors 16, 17 of each
robot's axis, the computer system 25 provides control signals C+ and C-
from the free wheels and signals F from the brakes of each robot's axis in
order to carry out the functions that are described hereinafter.
A robot according to the invention is designed so as to operate according
to two main modes, hereinafter referred to as "region mode" and "oriented
tracking mode". The "region mode" allows an operator to move a tool within
an area delineated by surfaces. The oriented tracking mode forces the
operator to closely follow a path in a predetermined direction.
A robot according to the invention can be used by a surgeon who has to
make, for example, a precise scalpel stroke. The scalpel stroke is then
stored as a path. During the approach phase toward the path, the surgeon
can first freely move the scalpel until a predetermined distance from the
patient; the robot is then used according to the region mode defining a
large authorized area. Then, the surgeon must engage the scalpel in a
funnel-shaped area that ends at the initial portion of the path. The path
can then be tracked according to the oriented tracking mode or the region
mode by defining a narrow channel area about the path.
FIGS. 2A and 2B illustrate two positions of a two-axis robot controlled
according to the region mode in order to force a point P of the robot to
stay in an area Z. Area Z is bidimensional and delineated by boundary
curves memorized in computer system 25. The robot includes a first axis
A.sub.1 according to the invention, articulating a first arm 41 on a
framework. A second arm 42 is articulated at the free end of arm 41 by a
second axis A.sub.2 according to the invention. Point P is disposed at the
free end of arm 42.
The computer system constantly calculates the position of extremity P by
using the information provided by the sensors of axes A.sub.1 and A.sub.2
as well as the lengths of arms 41 and 42 in accordance with the
above-mentioned relation P=f(A).
In FIG. 2A, extremity P is within the authorized area Z, and axes A.sub.1
and A.sub.2 are controlled so as to allow free rotation in both directions
S+ and S-. Thus, extremity P can freely move in any direction.
In FIG. 2B, extremity P moves beyond the boundary of area Z. Then, each
axis A.sub.1 and A.sub.2 is controlled so as to authorize a respective
rotation only in a direction (S- for both axes) adapted to urge extremity
P to penetrate into area Z.
With the above-described operation, an operator can move the extremity P
only within area Z; any attempt to move beyond this area being immediately
detected and stopped by the control of the appropriate free wheels of axes
A.sub.1 and A.sub.2.
A real control of the robot requires a response time (time for calculating
the positions and reactions of the disengageable free wheels) that does
not allow the robot to immediately detect whether the point P to be
controlled has reached a limit position and to instantaneously react.
Thus, depending on the moving speed of point P, the latter can, before
detection, move beyond the limit position. Here, the role of the speed
limiter 14 clearly appears. By limiting the speed to a known value, the
response time being known, the maximum overrun value is thus known.
Therefore, this overrun can be adjusted by regulating the limit speed.
For the sake of simplicity, FIGS. 2A and 2B represent a plane case. In a
real case, the boundary area Z is three-dimensional if it is desired to
check the position of a point, and is six-dimensional if it is also
desired to check the orientation of a tool.
Thus, more generally, the "region mode" consists in memorizing boundary
surfaces within which the tool is allowed to freely move. Preferably, the
boundary surfaces are divided into simple surfaces, such as planes,
cylinder portions, sphere portions, etc. The computer system 25 then
constantly calculates the minimum distance between the tool and a boundary
surface. If the tool reaches or moves beyond a boundary surface, the
robot's axes must be controlled so that the tool can be moved only
inwardly, with respect to the boundary surface.
If it is desired to control the movement of a tool according to n degrees
of freedom (n.ltoreq.6), the boundary surfaces are surfaces having a
dimension n-1 (hypersurfaces of an n-dimension space). Hereinafter, terms
such as "point", "position", "path", etc. relate to an n-dimension space,
that is, for example, a point defines a tool's position and orientation.
In order to control the various axes of the robot when it moves beyond a
boundary hypersurface, it can be proceeded as follows.
First, one determines the normal to the boundary surface that is defined as
the line intersecting point P corresponding to the current position of the
tool and point H of the boundary surface, the closest to point P. One
calculates from this normal line a vector N oriented toward the inside of
the boundary surface. Then, relation
A.sub.N =[J.sup.-1 ]N
is applied, where [J.sup.-1 ] is the reverse matrix of the above-mentioned
matrix [J], associating a position variation dP with a variation dA of the
articulation variables A.
Then, a set A.sub.N of rotation angle values is obtained;
if a value is positive, the rotation of corresponding axis is allowed only
in a positive direction,
if a value is negative, the rotation of corresponding axis is allowed only
in a negative direction,
if a value is zero, the corresponding axis is blocked.
The number n of degrees of freedom of the robot to be effectively
controlled according to the invention can be smaller than the total
number. In that case, the remaining degrees of freedom are blocked or
free, for example if it is desired to prevent or allow, respectively, a
continuous free rotation of the tool about an axis.
FIG. 3 shows how to proceed according to the oriented tracking mode to
follow a stored path T, also considered in an n-dimension space. The
following steps are carried out in an infinite loop:
calculating the current position P of the tool;
determining a vector N intersecting position P, perpendicular to the path,
and oriented toward the path;
determining a vector Tg orthogonal to vector N (tangent to the path) and
oriented in the desired direction;
calculating vector V which is a linear combination having positive
coefficients of vectors N and Tg; and
calculating a set A.sub.V corresponding to the rotation angle values of the
robot's axes by applying the reverse Jacobian matrix [J.sup.-1) to vector
V.
Thus, a set A.sub.V of angle values is obtained. One applies to this set of
values the above rules defined for set A.sub.N. As a result, the tool can
be moved, from the current position according to the desired direction
(Tg) toward the path (N).
The coefficients associating vector V with vectors N and Tg depend on the
distance between position P and the path. So, if the distance is
significant, the direction of vector V is close to that of vector N; if
the distance is small, the direction of vector V is close to that of
vector Tg.
According to a variant of the described methods allowing to reduce the
calculation time, the boundary surfaces and paths are previously stored in
the space of articulation variables. To achieve this purpose, formula
A=f.sup.-1 (P), where f.sup.-1 is the reverse function of the
above-mentioned function f, is applied to suitably selected points P of
surfaces and paths considered in the position and orientation space
(missing points can be interpolated, if required). Therefore, it is
unnecessary, in order to determine the rotation direction of the axes, to
calculate in real time matrices J and their respective reverse matrices.
The axis rotation directions are then directly provided by the signs of
the components of vectors N (for the region mode) and V (for the oriented
tracking mode), these vectors being determined in accordance with the
above-mentioned steps applied in the articulation variable space.
A robot according to the invention can also be used to enable an operator
to move a tool toward a precise position P.sub.0. A set of values a0.sub.j
of the rotation angles of the axes corresponds to position P.sub.0. The
rotation direction of each axis A.sub.j, as long as the position is not
reached, is determined by the sign of expression a0.sub.j -a.sub.j, where
a.sub.j is the current value of the rotation angle of axis A.sub.j. This
mode of utilization can be compared with the region mode whose authorized
area is limited to the point P.sub.0 to be reached.
In order to facilitate tracking of the path by the operator, he can be
provided with position information given in various ways (sound, display).
It is possible, for example, to display on a screen the stored path as
well as the current position of the tool. Additionally, in order to
provide a stronger guiding feeling, brakes 14 can be controlled so that
their braking force is proportional, for example, to the difference
between the position and the path.
A robot according to the invention can include an assembly of light arms
for ensuring the precise path and an assembly of strong arms for ensuring
the approach phase to the robot, the strong arms being blocked, if
required with respect to a device detecting the position of a patient,
once the approach phase is completed.
The control computer system 25 can be positioned with respect to the
patient with systems known per se, such as those already used in
neuro-surgery to position a cylinder for guiding a needle with respect to
a patient's head.
The invention has been described in relation with robots having rotation
axes. Of course, it similarly applies to robots including translation
axes.
Those skilled in the art will be able to program the computer system 25 in
order to suitably control the axes according to the invention as a
function of the calculation of the position and the stored path to be
complied with. They will also be able to select or fabricate disengageable
free wheels.
As is apparent to those skilled in the art, various modifications can be
made to the above disclosed preferred embodiments. They will be able to
find many control methods equivalent to the methods described. For
example, to track a path T, from the position shown in FIG. 2A, one of the
axes (for example A.sub.1) can be blocked and another axis (A.sub.2) can
remain free in order to move extremity 43 in an arc of a circle up to
limit T2. When limit T2 is reached, axis A.sub.2 is blocked and axis
A.sub.1 is released so as to move extremity 43 up to limit T1 where axis
A.sub.1 is blocked again and axis A.sub.2 is released.
Various applications, not described, can be found for a robot according to
the invention, especially in the reeducation field, for motion learning,
etc., the described tool being then, for example, a strap fixed to a limb.
* * * * *
|
|
 |
|
 |
Description |
|
