Image-directed robotic system for precise robotic surgery including redundant consistency checking

Description Submit all comments and votes

FIELD OF THE INVENTION

This invention relates generally to robotic systems and, in particular, to a robotic system that integrates an interactive Computed Tomagraphy (CT)-based presurgical planning component with a surgical system that includes a multiple-degree of freedom robot and redundant motion monitoring. An illustrative application is presented in the context of a system that prepares a femoral cavity to have a shape precisely determined for receiving a cementless prosthetic hip implant.

BACKGROUND OF THE INVENTION

It has been found that computed tomagraphy (CT) imaging and computer modelling methods provide a precision for pre-surgical planning, simulation, and custom implant design that greatly exceeds the precision of subsequent surgical execution. For example, approximately one half of the 300,000 total hip replacement operations performed each year use cementless implants. Stability of the implant, uniform stress transfer from the implant to the bone, and restoration of the proper biomechanics critically affect efficacy and, in turn, are significantly affected by the proper placement of the implant relative to the bone. An important factor in achieving proper placement of the implant is the accuracy with which the femoral cavity is prepared to match the implant shape.

Recently reported research confirms that gaps between implant and bone significantly affect bone ingrowth. One study of the standard manual broaching method for preparing the femoral cavity found that the gaps between the implant and the bone is commonly in the range of one millimeter to four millimeters and that the overall resulting hole size was 36% larger than the broach used to form the hole. As a result, only 18-20 percent of the implant actually touches bone when it is inserted into such a hole. Furthermore, the placement of the implant cavity in the bone, which affects restoration of biomechanics, is as much a function where the broach "seats" itself as of any active placement decision on the part of the surgeon.

Typically, precise surgical execution has been limited to procedures, such as brain biopsies, for which a suitable stereotactic frame is available. However, the inconvenience and restricted applicability of these devices has led some researchers to explore the use of robots to augment a surgeon's ability to perform geometrically precise tasks planned from CT or other image data.

Safety is an obvious consideration whenever a moving device such as a robot is used in the vicinity of a patient. In some applications, the robot does not need to move during the "in-contact" part of the procedure. In these applications the robot moves a passive tool guide or holder to a desired position and orientation relative to the patient. Brakes are then set and motor power is turned off while a surgeon provides whatever motive force is needed for the surgical instruments. Other surgical applications rely on instrumented passive devices to provide feedback to the surgeon on where the instrument is located relative to an image-based surgical plan.

In an IBM Research Report (RC 14504 (#64956) 3/28/89) R. H. Taylor et al. describe a robotic system for milling a correctly shaped hole into a femur for receiving a cementless hip implant. The system computes a transformation between CT-based bone coordinate data and robot cutter coordinates. The transformation is accomplished in part by a combination of guiding and tactile search used to locate a top center of each of three alignment pins that are pre-surgically affixed to the femur and CT-imaged. This robotic system includes a vision subsystem to provide a redundant check of the robot's motion to ensure that the tool path does not stray outside of a planned work volume. An online display is provided for the surgeon to monitor the progress of the operation. Proximity sensors may be positioned to detect any subsequent motion of the pins relative to a robot base.

SUMMARY OF THE INVENTION

The invention discloses a robotic surgical system that includes a multiple degree of freedom manipulator arm having a surgical tool. The arm is coupled to a controller for controllably positioning the surgical tool within a three dimensional coordinate system. The system further includes apparatus for determining the position of the surgical tool in the three dimensional coordinate system relative to a volumetric model. The determining apparatus includes a device for detecting a location of the surgical tool, such as an optical tracking system disposed for imaging a region of space that includes at least a part of the manipulator arm. An output of the tracking system is coupled to a processor that processes the volumetric model for determining if the surgical tool is positioned outside of a predetermined volume of space. The system further includes redundant safety checks including a strain gage for detecting in three dimensions any slippage between an immobilized tissue, such as bone, and a reference point and also a force sensor coupled to the surgical tool. The multiple and redundant safety devices are employed for suspending a motion of the surgical tool to prevent the tool from operating outside of the volumetric model. The coordinates and structure of the volumetric model are determined during a pre-surgical planning session wherein a surgeon interactively selects and positions a suitably shaped implant relative to images of the bone within which the implant is to be installed.

BRIEF DESCRIPTION OF THE DRAWING

The above set forth and other features of the invention are made more apparent in the ensuing. Detailed Description of the Invention when read in conjunction with the attached Drawing, wherein:

FIG. 1 is a block diagram showing a presently preferred embodiment of a surgical robotic system;

FIG. 2a illustrates a human hip prosthetic implant;

FIG. 2b illustrates the implant of FIG. 2a implanted with a femur, the Figure also showing the placement of three alignment pins upon the femur;

FIG. 3a illustrates a method of determining a cutter work volume for the implant of FIG. 2a;

FIG. 3b is a CSG tree representation of the cutter work volume of FIG. 3a;

FIG. 4 is a graph of measured force, resolved at the cutter tip, as a function of time, the graph further illustrating first and second force thresholds;

FIG. 5 illustrates information flow during pre-surgical planning; and

FIG. 6 is a block diagram showing in greater detail the pre-surgery system and the coupling of the pre-surgery system to the robot controller and on-line display.

DETAILED DESCRIPTION OF THE INVENTION

FIG. 1 is a block diagram illustrating a present embodiment of a surgical robotic system 10. System 10 includes a robot 12 having a manipulator arm 14 and a base 16. The arm 14 has at least 5-axes of motion and includes a plurality of joint (J) motors J1, J2, J3 and J4 that provide four degrees of motion and, in the present embodiment, a pitch motor (PM) 18 that provides the fifth degree of motion. In the present embodiment of the invention the robot 12 is a four-degree of freedom IBM 7575 or 7576 SCARA manipulator having an additional pitch axis (IBM is a registered trademark of the International Business Machines Corporation). The PM 18 is presently implemented with a stepper motor having an encoder, primarily for stall detection, that is controlled with a commercially available indexer of the type that accepts a user provided acceleration, velocity and distance commands. Each of the manipulator arm joints has an associated controlling microprocessor.

The robot 12 further includes a six degree-of-freedom wrist-mounted force sensor 20. One suitable force sensor is known as a Lord Force/Torque Wrist Sensor, Model F/T-30/100 having a maximum force limit of 30 pounds and a resolution of 1/40 pound. The force sensor 20 is coupled via a serial or a parallel interface to a robot controller 24.

The robot 12 further includes an end effector having a cylindrical high-speed (65000 rpm) pneumatic surgical cutting tool 22. During surgery all but the robot's end-effector are covered by a sterile sleeve, the end-effector being separately sterilized. The robot 12 is positioned relative to an operating table such that it has ready access to the surgical region. The robot controller 24 provides servocontrol, low-level monitoring, sensor interfaces, and higher-level application functions implemented in the AML/2 language. The controller 24 is presently embodied in an industrial IBM Personal Computer AT data processor herein low-level servo control and force sensor interface is provided by suitable printed circuit cards that are plugged into the processor bus (Personal Computer AT is a registered trademark of the International Business Machines Corporation). Controller 24 includes an AML/2 Language Interpreter and also Motion Control System (MCS) software. A commercial version of this software is described in "AML-2 Language Reference Manual", Manual #G7X1369 and in "AML-2 Manufacturing Control System User's Guide", Manual #G67X1370, both of which are available from IBM Manufacturing Systems Products, Boca Raton, Fla.

In the present embodiment the AML/2 software is modified to accommodate the operation of the PM 18 and the force sensor 20. During surgery, the force sensor 20 is employed in conjunction with a Force Monitoring Processor (FMP) 53 to support redundant safety checking, tactile searching to locate aligning pins, and compliant motion guiding by the surgeon.

The FMP 53 is interfaced to the wrist-mounted force sensor 20 and computes forces and torques resolved at the utter 22 tip. As can be seen in the graph of FIG. 4 if any cutter 22 tip force component greater than approximately 1.5 kgf (L.sub.1) is detected, the robot controller 24 is signalled to freeze motion. Forces greater than approximately 3 kgf (L.sub.2) result in arm power being removed, an arm 14 "shutdown" condition. The force sensor 20 is effective in detecting such conditions as the cutter 22 stalling, encountering improperly excised soft tissue, and changes in bone hardness such as that which occurs between trabecular and cortical bone.

Also coupled to the controller 24 is a hand-held pendant 26 for use by the surgeon as an input terminal as will be described.

A motion monitoring subsystem includes, in the present embodiment, an optical tracking system having a camera 28 with three spatially separated image sensors 30a, 30b and 30c. Coupled to camera 28 is a camera processor 32 that visually tracks in three dimensions the position and orientation of a plurality of infrared (IR) beacons, such as LEDS 34, that are mounted on a reference plate 36 coupled to the robot end effector. The optical tracking system is but one of several redundant motion detecting systems employed during the cutting phase of the surgery to verify that the cutter 22 tip does not stray more than a specified amount outside of a defined implant volume. In a presently preferred embodiment of the invention the optical tracking system is a type known as Optotrak that is manufactured by Northern Digital, Inc. The three image sensors 30a-30c are line-scan devices mounted in a rigid frame which track the position of IR LEDs 34 in three dimensional space to an accuracy of 0.1 mm/m.sup.3.

In the present embodiment the reference plate 36 includes eight LEDs 34 which function as positional beacons. Camera processor 32 software is employed to compute a camera-based coordinate system from the beacon locations. Robot-to-camera and cutter-to-reference plate transformations are computed by a least squares technique from data taken with the robot arm 14 fn various known positions, using appropriate linearized models.

An output 32a of the optical tracking system is coupled to a safety monitoring processor 38 that has as one function a task of verifying that the cutter 22 remains within the predetermined spatial volume associated with the selected implant. Processor 38 receives a coordinate transformation (Tcp) of the reference plate 36 relative to camera 28 from the camera processor 32. In an alternative embodiment, processor 38 receives the coordinates of LEDS 34 from camera processor 32 and computes the coordinates of reference plate 36 itself.

During a calibration phase prior to surgery, robot controller 34 moves the robot to a plurality of positions and orientations. After each motion, robot controller 24 transmits the position and orientation of the robot's cutter 22 to processor 38 over a communication bus 24a. Processor 38 uses this information, together with the reference plate 36 coordinates relative to the camera 28, to compute the coordinate transformation (Trc) between the camera coordinate system and the robot coordinate system and also to compute the coordinate transformation (Tpk) between the cutter 22 and the reference plate 36. As a result, processor 38 is enabled to determine the coordinate transformation (Trk) of the robot's cutter 22 relative to the robot from the relationship

Trk=Trc*Tcp*Tpk. (1)

In the presently preferred embodiment, the calibration for Tpk is accomplished by placing plate 36 in many orientations with respect to camera 28 with the tool tip being maintained in the same location. This may be accomplished either by reliance on the robot-to-tool calibration or preferably by means of a tactile search procedure using force sensor 20 to locate the cutter tip at a known constant position relative to a calibration pin or post 54. Alternative embodiments include the use of other sensing means either for direct measurement of the cutter tip position or as feedback allowing the robot controller 24 to place the cutter tip in a known position relative to a reference landmark or coordinate system.

In other embodiments of the invention the position of the manipulator arm 12 may be tracked in three dimensions by, for example, magnetic sensing devices or by an ultrasonic ranging system. That is, the practice of the invention is not limited to use with an arm motion detector that relies on detecting optical beacons.

During surgery processor 38 receives inputs over communication bus 24a from the robot controller 24 specifying, for this embodiment, a Constructive Solid Geometry (CSG) tree representation of the volume to be checked ("check volume") and the coordinates of this volume relative to the robot 12. Processor 38 repeatedly receives reference plate coordinates (Tcp) from camera processor 32 or, alternatively, receives the coordinates of LEDS 34 and computes (Tcp). Processor 38 computes Trk and determines if the cutter 22 is within the specified check volume.

Processor 38 also monitors a bone slippage detector which, in accordance with an aspect of the invention, is comprised of strain gages 40 which are physically coupled to a tissue, such as a bone 42, that is being surgically altered. The strain gages 40 are disposed to measure in three dimensions any displacement of the bone 42 relative to a bone fixator 44, the fixator 44 being rigidly coupled to the robot base 16. The strain gages 40 are interfaced through appropriate circuitry, including an analog-to-digital converter, to the processor 38.

It has been demonstrated that motions on the order of 0.1 mm are readily detectable in this manner. Furthermore, it has been determined that with a suitable fixator 44, such as a device that employs screw-type connections made directly to the bone, that even rather large forces (5 kgf) produce only a few microns of motion. Bone motion of this small magnitude is negligible in the context of this application. Thus, although no significant bone motion is expected during surgery, the strain gages 40 provide an immediate indication if any bone slippage should occur.

If slippage of the bone is detected at least two options are available. A first option is to recalibrate the system by relocating the position of the bone in space by locating the three pins 46 with the robot effector in accordance with a procedure described below. A second option is employed if the slip sensor is accurately calibrated. The second option involves a mathematical determination of the amount of bone slippage to derive a compensation factor that is applied to subsequent robot motions. The first option is preferred for simplicity.

Further in accordance with the invention, and as illustrated in FIGS. 3a and 3b, the safety monitoring processor 38 employs a volumetric processing technique to verify that the cutter 22 tip does not stray by more than a predetermined tolerance beyond a spatial envelope that corresponds to the three dimensional implant model. The present embodiment of the invention employs the above mentioned CSG tree "check volumes", corresponding to shapes resulting from implant and cutter selection, that are constructed from primitives bounded by quadric surfaces located at a defined distance, such as one millimeter, outside of the furthest nominal excursions of the cutter 22. In FIG. 3a the volume of space that corresponds to a selected implant shape is determined by partitioning the implant shape into a plurality of primitive shapes that correspond to (a) a cutter approach volume, (b) an implant proximal portion and (c) an implant distal portion. The inner dashed line corresponds to the maximum cutter 22 excursion as measured from the center (X) of the cutter 22. The cutter 22 outer edge is thus coincident with the outer, solid envelope of the implant volume. In practice, the " cutter center (X) is uniformly offset away