Field Operation of a Surgical Robot via Airborne Wireless Radio Link

Transcription

1 Field Operation of a Surgical Robot via Airborne Wireless Radio Link M.H. Lum, D.C.W. Friedman, H.H. King, Timothy Broderick, M.N. Sinanan, J. Rosen and B. Hannaford Dept. of Electrical Engineering, University of Washington, {mlum, hawkeye1, rosen, blake}@u.washington.edu Dept. of Mechanical Engineering, University of Washington dwarden@u.washington.edu Dept. of Surgery, University of Cincinnatti brodertj@ucmail.uc.edu Dept. of Surgery, University of Washington mssurg@u.washington.edu Abstract Robotic assisted surgery generates the possibility of remote operation between surgeon and patient. We need better understanding of the engineering issues involved in operating a surgical robot in remote locations and through novel communication links between surgeon and surgery site. This paper describes two recent experiments in which we tested a new prototype surgical robot manipulation system in field and laboratory conditions. In the first experiment, we set up the robot in a remote pasture land and ran it on generator power. Telecommunication with the surgical control station was provided by a novel airborne radio link supported by an unmanned arial vehicle. In the second experiment, we teleoperated the robot over an Internet link between Imperial College London and our laboratory in Seattle. Data are reported on surgeon completion times for basic tasks and on network latency experience. The results are a small step towards teleoperated surgical robots which can be rapidly deployed in emergency situations in the field. I. INTRODUCTION The primary objective of this project was to demonstrate in the field a Mobile Robotic Telesurgery (MRT) system which would eventually allow a remote surgeon to operate on a patient regardless of their location or environment. Remote environments limit access to power and telecommunication resources needed by telesurgery systems. Our experiments tested a novel MRT system comprised of a prototype next generation surgical robot and an Unmanned Airborne Vehicle (UAV)-based communication system in range land near Simi Valley, California. Network limitations in latency, bandwidth, jitter, packet loss and loss of signal make telesurgical intervention in extreme environments difficult. While other existing topologies such as geosynchronous satellites can provide wireless communication for use in mobile telesurgery, long latency as the result of signal transmission distance precludes robust use of satellite communications in telesurgery. Geosynchronus orbit is 35,900 km above the earth surface, and a complete loop from surgeon to robot and back requires two trips up and down; a 0.48 sec round trip at the speed of light. Various types of Unmanned Airborne Vehicles can operate at altitudes from 100 to 20,000 meters and carry equipment to provide robust, low-latency, high quality communications. In these experiments, the communication link was provided by a single bounce of the signal off a Puma UAV by AeroVironment Inc., Simi Valley, CA. The Puma is a 2-meter wingspan, Small Unmanned Aerial Vehicle (SUAV) that flies at altitudes below 5000 meters above sea level and can provide line of sight communication up to a distance of 12 kilometers with low gain antennas and 20km with higher gain antennas. Approximately 4,000 Pumas are currently deployed across the world in military applications. A. Literature Review The earliest teleoperators, invented by Ray Goertz of Argonne National Labs in the late 1940 s and early 1950 s,[1] are mechanical hands, coupled through a system of cables and pulleys to a remote handle which the operator can control at a safe distance. These mechanisms were very effective and are still in use, but are limited by their mechanical nature to short distances and similar size scales. After continuous development at many labs[2], [3], teleoperation technology has matured enough in the last 15 years to apply to surgery. By connecting the surgeon to the tools through teleoperators, the tools can be made smaller than the human hand, more dexterous in small body cavities, and, perhaps most revolutionary, can connect surgeon and patient across large distances. In the early 1990 s Dr. Phil Green at SRI International developed a two-handed teleoperated surgery unit for DARPA[4]. This highly influential project encouraged the startup of two companies to address the civilian surgery market, Computer Motion Inc, of Goleta Ca., and Intuitive Surgical Inc. (ISI) of Silicon valley. Both of these companies developed FDA approved surgical robot systems, the Zeus from Computer Motion[?] and Da-Vinci from ISI[5]. In 2003 they merged under the name of ISI. Over 300 Da-Vinci systems are in use around the world today. Both of these systems are teleoperators, but neither had the capability to separate the surgeon and patient by more than a few feet.

2 In 2001, Dr. Jacques Marescaux worked with Computer Motion to develop a specially modified Zeus system and to perform the first remote surgical procedure (a gall-bladder removal, laparoscopic cholycystectomy) on a human patient[6], [7]. Dr. Marescaux controlled the robot from New York, and the patient was in Strasbourg France. Using a similarly modified Zeus, Dr. Mehran Anvari of McMaster University, has treated about 25 patients in Northern Canada from his offices in southern Ontario[8]. Recently, Broderick has evaluated remote surgical technology in extreme environments. With support from the US Army and NASA, he recently completed the NEMO mission a two week stay in the Aquarius underwater habitat, maintained by the US National Oceanographic and Atmospheric Administration (NOAA), 19 meters underwater off the Florida Keys. An updated version of SRI s manipulator was deployed in this cramped habitat and Dr. Anvari successfully operated the robot from Canada with up to 2 seconds of time delay[9], performing simulated surgical procedures and the handling of simulated lunar rocks 1. A major user of remote surgical technologies could be the military. Can they reduce or ameliorate combat casualties by getting the surgical care to the soldier faster than getting the soldier to the care? Surgeons are a scarce resource in the Military and with remote surgical technology they could be rapidly deployed right to where they are needed, even switch electronically from one battlefield to another in seconds even without the need to scrub in. The need for small and deployable surgical telerobotic system led to recent development of a new surgical robot by the authors at the University of Washington [10]. II. METHODS Detailed discussion of the robot s design are reported in [10]. The slave side of the surgical robot system (Figure 1) consists of two surgical manipulators that are positioned over the patient by a passive macro-positioner referred to as the C-arm. Design. The design specifications for the 7-DOF surgical manipulator were derived from extensive database of force, torque, and displacement measurments taken during animal surgery simulating human surgical procedures[11]. The robot mechanism is broken into three main pieces; the quasi-static base that holds up to seven actuators, the spherical mechanism that positions the tool, and the tool interface. The motion axes of the surgical robot are: 1) Shoulder Joint (rotational) 2) Elbow Joint (rotational) 3) Tool Insertion / Retraction (linear) 4) Tool Rotation (rotational) 5) Tool Grasping (rotational) 6) Tool Wrist-1 Actuation (rotational) 7) Tool Wrist-2 Actuation (rotational, optional) The first four joint axes intersect at the surgical port location, creating a spherical mechanism that allows for tool 1 missions/neemo9/default.asp Fig. 1. Photograph of the surgical robot system deployed in office space at the beginning of the HAPSMRT mission. manipulation similar to manual laparoscopy. DC brushless motors mounted to the base of the micromanipulator actuate all motion axes. Maxon EC-40 motors with 12:1 planetary gearboxes are used for the first three axes, which see the highest loads. Maxon EC-32 motors are used for the remaining axes. DC brushed motors are simpler to wire and control. However, brushless motors offer better torque to weight ratio and more efficiently dissipate heat because the windings are thermally coupled to the outer case of the motor. While the performance benefits of brushless motors are clear, they required more complex and expensive controllers and bulkier wiring. Each brushless DC motor (and associated optical encoder) required 14 conductors between motor and control box. The motors are mounted on quick-change plates, which allow for motor removal without the need for disassembling the cable system. The first two axes have power-off brakes to prevent tool motion in the event of a power failure. The cable system is comprised of a capstan on each motor, a pretension adjustment pulley, various pulleys to redirect the cables through the links, and attachment to each motion axis. The shoulder axis is terminated on a single partial pulley. The elbow axis has a dual-capstan reduction stage terminating on a partial pulley. The tool insertion / retraction axis has direct terminations of the cables on the tool holder. The tool rotation, grasping and wrist cables are terminated on capstans on a quick-change tool interface. The cable system transmission ratios (motor:joint) are: Shoulder: 7.7:1 Elbow: 7.3:1 and the insertion axis ratio is 133 radians per meter. Each axis is controlled by two cables, one for motion in each direction, and these two cables are pretensioned against each other. The cables are each terminated at both ends, to prevent any possibility of slipping. The cable system maintains constant pretension on the cables through the entire range of motion, however there are force and motion couplings between the axes, which must be accommodated by the control system. The mechanism links are machined from aluminum, and are generally I-section shapes with structural covers. These removable covers allow for access to the cable

3 system, while improving torsional stiffness of the links when they are in place. The links are also offset from the joint axis planes, allowing for a tighter minimum closing angle of the elbow joint. When used in laparoscopic procedures, laser pointers attached to the shoulder and elbow joints allow for visual alignment of the manipulator relative to the surgical port. When the two dots, visible on the abdomen, converge, the manipulator is positioned such that the center of rotation of the surgical manipulator is aligned with the pivot point on the abdominal wall. Tool Interface The tool interface controls the tool rotation, grasp, and wrist axes, and allows for quick changing of tools. A tool locks into position when pushed up the main axis until a snap is heard. Two parallel surfaces on the tool must be displaced together about 2mm to release the tool. This tool interface mechansim was designed so that a single human hand or a parallel jaw robotic grasper could remove or attache the tool with a simple grasping and pulling/pushing motion. The tools used are modified Micro-Wrist tools from the Computer Motion Zeus robot adapted to our tool interface. The tools grasp and wrist axes are actuated by pushrods in the tool shaft. High pitch threads are used to convert the rotational motion of the cable system capstans into linear motion of the tool pushrods. Because the modified Zeus tools have only one wrist axis, the final axis of our surgical robot is currently unused. The 7-DOF surgical manipulator must be positioned such that its center of rotation is coincident with the pivot point of the MIS port and oriented such that the workspace of the robot is aligned with the target anatomy. This is achieved with a 6- DOF positioning system that is supported by a frame structure which is rigidly affixed to the center column of a commercial motorized OR table but still allows tilt motions of the table. The first stage of the positioner is a linear lead-screw actuator from THK America that runs the length of OR table (1.6 meters motion range), one for each side. Next a passive 2-DOF planar mechanism with a Gitzo G1570M tripod head, referred to as the C-arm, positions and orients the surgical manipulator over the patient. At this point human assistants position the device in the vertical plane to align the mechanism to the port, and then manually lock the C-arm. An actuated C-arm has been designed and future effort will focus on development, fabrication and integration of this piece of hardware into the complete system. A. Software and Controls The critical link between the controls software, running on a RTAI Linux computer, and the motor controllers is a specially designed USB 2.0 interface board. Our USB board features eight channels of high-resolution 16bit digital to analog conversion for control signal output to each controller and eight 24bit quadrature encoder readers for position feedback. The board is a general purpose low-level interface between any generic PC and any robotic device with optical encoder feedback. Our Linux control software can read the sensors and write new output commands to all 8 channels in 120µsec. When surgeons transitioned from open surgery to MIS, the sense of touch was greatly diminished. Friction between the surgical tool and the seal of the MIS port, as well as the forces exerted on the port by the abdominal wall greatly reduce the surgeons ability to feel the tissue they interact with. With the right master interface device, surgical robots present an opportunity for surgeons to regain and potentially enhance their sense of touch in MIS procedures. One of the design goals for our robot was to make it lightweight and back driveable to create a slave that is force-feedback capable for bilateral teleoperation, however at this stage we have not yet implemented or tested force feedback capabilities. Master Console. As an interim measure prior to development of a full surgery console, we have developed a low cost and highly portable surgical control station from off-the shelf components. Remote users need only install a software CD to connect to our robot. Currently our system uses two PHANToM Omni s (SensAble Inc.) as the master device, one for each arm. The Omni is a cost effective solution that allowed us to quickly implement a surgeon interface device for our master/slave system. It features 3-DOF force-feedback and 6- DOF sensing. Our master console also features a foot-pedal that enables and disables the virtual coupling between the Omni and the surgical manipulator. This will allow for position indexing as well as allowing the surgeon to put down the Omnis stylus without moving the robot. Future development our system will include a specialized force-feedback surgeon console and interface specifically designed for high quality multisensory feedback in the surgical context. B. Software and Safety Architecture The control system and surrounding electronic hardware were designed to incorporate safety, intelligence, modular design, and flexibility[10]. As this is a medical device, the most critical of these aspects is safety. We have not yet designed the system to a full human-rated standard of safety. However, we paid careful attention to safe and reliable operation despite software malfunctions which are extremely difficult to completely eliminate during system development and early animal testing. To achieve this reliability, we defined four well-defined states and a basic set of transition rules for the hardware and software system (Figure 2). A programmable logic controller (PLC) controls all transitions between these states. The system states include: Initialization, Pedal Up, Pedal Down, and Emergency Stop. The Pedal Up state is initiated when the surgeon lifts his/her foot from the foot-pedal and decouples the master from the surgical manipulator. This is done to pause, to perform tool indexing, or to free the surgeons hands for peripheral tasks. The Pedal Down state is initiated when the surgeon pushes the foot-pedal down allowing the master device to directly control the surgical manipulator. C. Engineers Interface The Engineering Interface (EI), a low-level interface to the states and mechanisms of the control software, assists

4 north of Simi Valley, CA for telesurgery experiments on an inanimate model. Fig. 2. Surgical robot Control System State Diagram. Software can only change state in response to a change in the PLC state. Software can request state change from the PLC. External input from E-Stop physical circuit, or failure to detect a 50Hz watchdog signal from the software will trigger an E-stop state from any state. robot development. Developers are presented with a graphical user interface (GUI) with easy access to robot features. In development stages, the four system states (Figure 2) can be set manually with the click of a button. Control commands can be sent to any degree of freedom or the entire robot for example, a 40 degree sine wave can be output on the shoulder joint, motor controller number two can be made to output 30% maximum current, or endpoint position can be instructed to move 3cm left in Cartesian space. Robot variables (motor output, joint position, end-effector position) are displayed on-screen in real-time, and also logged for later evaluation. The EI can connect to the RTAI Linux control system by two FIFO device nodes or a single, bi-directional (TCP/IP) network socket. Two types of data are exchanged: a packet containing all robot-state information is received by the EI, and a command packet with all instruction parameters is sent from the EI to the control software. This link is independent of the master-slave link. III. TELESURGERY FIELD EXPERIMENT Many research systems are conceived, developed and tested in a lab environment. To fully demonstrate the applicability of a system it must be deployed into the environment for which it was designed. Dr. Timothy Broderick, MD of the University of Cincinnati led the HAPs/MRT project to evaluate surgical robotics in field conditions. In early June of 2006, the surgical robot system was taken from our lab and deployed a few km A. Operations in Remote Location The fielded system consisted of two stations connected by a wireless data link relayed by UAV (Figure 3). The system was powered by construction grade internal combustion generators and was set-up under portable tents in an isolated field. Separated by a distance of 100 meters, the surgical manipulators and the master control console were connected via a wireless digital datalink relayed by AeroVironment s PUMA unmanned aircraft up to 1 km away at an altitude of about 100 meters. The datalink provided by AeroVironment utilized Internet Protocol communication at a rate of 1MB per second between the two sites allowing the master/slave communication architecture to remain unmodified for this experiment. A single NTSC color video signal was provided using a digital video camera which acquired a closeup picture of the surgical site. A second video channel was sometimes used to display the overall operating environment. HaiVision Inc. (Montreal, Canada) provided a hardware codec that transmitted the video signal at 800kbps in MPEG-2 format. Drs. Broderick and Huffman performed a series of tasks including touching a series of landmarks and suturing on a latex glove stretched over a box (Figure 4). The gloved box was marked with a circle, and a grid of 10 landmarks spaced 1cm apart left to right and 0.5cm apart toward and away. The landmarks were numbered 1-10 starting with 1 at the upper left, and finishing with 10 at the lower right. The following five tasks were part of the experimental protocol. 1) Right hand touch each landmark in numeric order 2) Left hand touch each landmark in numeric order 3) Touch each landmark in numeric order using alternating left and right hands. Right hand touches the odd numbered landmarks, left hand touches the even numbered ones 4) Right hand trace inner edge of circle clockwise 5) Left hand trace inner edge of circle clockwise During three days of field deployment, kinematic data of the surgeons commands and the surgical manipulators motions were collected along with network characterization data. Control information from the master side was transmitted at 100 packets per second in the field trial and 1000 packets per second in a later transcontinental trial (below). B. Transcontinental Experimment In collaboration with Julian Leung, George Mylonas, Sir Ara Darzi and Ghuang-Zhong Yang from Imperial College (London, England) we demonstrated the ability to operate across a long distance. In the lab in London, two PHANToM 6- DOF Premium haptic devices were equipped with our surgeon console software. ichat software (Apple Computer Inc) was used for close-up video feedback of the surgical site. Skype video was used for transmision of operating room video. Drs.

5 Configuration Network (ms) Video (ms) Wireless (UAV) Wired (Internet) TABLE I TIME DELAYS IN TWO TELEROBOTIC SURGERY CONFIGURATIONS. Subject Exp F F I I TABLE II TASK COMPLETION TIMES (SECONDS) FOR TWO SURGEONS EACH IN THE TWO EXPERIMENTS. Darzi and Leung, performed the same experimental protocol described above. Time delays for the two test configurations are given in Table I. In the field experiment, with the UAV and HaiVision devices, the delays were substantially shorter. Internet latency (measured by ping) from Imperial College to our Lab in Seattle, Washington USA, was about 140 ms and ichat video encoding/decoding delay was about 1 second. This experiment showed both that the master console software was flexible enough to adapt to other PHANToM devices, and the ability of the system to teleoperate across long distances. Task performance data for the four surgeons in two experiments are collected in Table II. The variations between individual surgeons are pronounced compared to those between the two experimental conditions: field trials ( F ) and Intercontinental Internet link ( I ). Further experiments must be performed to generate statistically significant results. Two example trajectories from one of the surgeons for task 1 from the Intercontinental Internet experiments are plotted in Figure 5. The numeric data has yet to be completely processed and will be the subject of future reports. IV. CONCLUSIONS We have reported initial tests of a surgical robot in the field and when controlled over a global Internet link. Preliminary data indicate that the robot can be effectivley controlled in basic positioning tasks. The rigors of the field test exposed some weaknesses such as low generator voltage output (which automatically shut down our power conditioners) and poor control of tool orientation due to mechanical interferences inside the mechanism (which has now been fixed). For some applications, the 1-10km range of the low altitude UAV will limit applicability. We plann future MRT experiments using a High Altitude Long Endurance (HALE) UAV, also from Aerovironment, to provide ranges on the order of 10-50km. A novel feature of our current system is a low cost surgeon user interface built entirely from commercially available hardware. In our field trial, we obtained very high video quality using dedicated codec hardware. In our intercontinental trial, we substituted freely available video chatting software with surprisingly good results. Although it is unlikely that basic technology like ichat video will statisfy clinical surgeons, it provides a very convenient platform for experimentation. There are many more engineering steps required before a system such as ours can be deployed for actual use and we are planning future experiments to address them. For example, redundancy should be built into the communication link in some form. We feel that none of these barriers are insurmountable. Deploying the system into a field environment and successfully executing the experimental protocol demonstrated the feasibility of performing Mobile Robotic Telesurgery (MRT) in a remote environment and with long communication links. ACKNOWLEDGEMENTS The authors would like to acknowledge support from the US Army TATRC program, and our colleagues at The University of Cinncinnati, AeroVironment, Imperial College London, and HaiVision Inc. We also acknowledge significant technical contributions from Regina Donlin, Clint Bland, and Levi Miller. REFERENCES [1] R.C. Goertz and W.M. Thompson. Electronically controlled manipulator. Nucleonics, pages 46 47, November [2] A. K. Bejczy. Sensors, controls, and man-machine interface for advanced teleoperation. Science, 208: , June [3] A.K. Bejczy and Z. Szakaly. Universal computer control systems (uccs) for space telerobots. Proceedings IEEE International Conference on Robotics and Automation, 4, Mar [4] J.W. Hill, P.S. Green, J.F. Jensen, Y. Gorfu, and A.S. Shah. Telepresence surgery demonstration system. Proceedings., 1994 IEEE International Conference on Robotics and Automation, 3, 05/08/ /13/1994. [5] G.S. Guthart and J.K. Salisbury. The intuitive tm telesurgery system: Overview and application. Proceedings IEEE International Conference on Robotics and Automation (ICRA), April, [6] J. Marescaux. Transatlantic robot-assisted telesurgery. Nature, 413, Sept. 27. [7] J. Marescaux, J. Leroy, F. Rubino, M. Smith, M. Vix, M. Simone, and D. Mutter. Transcontinental robot-assisted remote telesurgery: Feasibility and potential applications advances in surgical technique. Annals of Surgery, 235(4), April [8] A. Pirisi. Telerobotics brings surgical skills to remote communities. The Lancet, 361(9371). [9] M. Anvari, T. Broderick, H. Stein, T. Chapman, M. Ghodoussi, D. W. Birch, C. Mckinley, P. Trudeau, S. Dutta, and C.H. Goldsmith. The impact of latency on surgical precision and task completion during robotic-assisted remote telepresence surgery. Computer Aided Surgery, 10(2), March [10] M.J.H. Lum, D. Trimble, J. Rosen, K. Fodero, H. King, G. Sankarayanaranan, J. Dosher, R. Leushke, B Martin-Anderson, M.N. Sinanan, and B. Hannaford. Multidisciplinary approach for developing a new minimally invasive surgical robot system. In Proceedings of the 2006 BioRob Conference, Pisa, Italy, February [11] J. Rosen, J.D. Brown, L. Chang, M. Barreca, M. Sinanan, and B. Hannaford. The bluedragon - A system for measuring the kinematics and the dynamics of minimally invasive surgical tools in vivo. In Proc. IEEE Intl. Conf. on Robotics and Automation ICRA-2002, pages , Arlington VA, May 2002.

6 Fig. 3. Telesurgery experimental setup. (top) Block diagram of the experiments with UAV wireless link and direct internet connection options. (bottom) Panoramic photo of the wireless experimental site. Control station at left, robot at right. UAV is visible in center of photo on ground.

7 Fig. 4. (a) Experimental protocol was performed on a rubber glove stretched over a small box (b) Successful suture tied on glove. Fig. 5. (left) Motion in the X-Y plane from one surgeon performing Task 1 (see text) in the intercontinental Internet link experimment. Time function of Y-axis date during same movement (right).