Surgical Master Console Using General-Purpose Robot Arms and a Separable Articulated Distal Interface: Porcine In-Vivo Evaluation

I. INTRODUCTION

• Dedicated surgical masters are costly and difficult to reproduce, while accessible haptic devices do not directly provide an articulated surgical wrist with continuously variable grasp.Robot-like leaders typically rely on close leader–follower kinematic correspondence, which is unavailable for a general-purpose arm paired with a non-isomorphic surgical follower.Source: §I · p. 1
• Hand-guided control characterizes current scaling, gravity, break-away friction, and thermal loading to support gravity compensation and motion-gated assistance during sustained use.The controller and endpoint-independent master-state interface were used with simulation and an RCM-constrained dual-FR3 platform.Source: §I · p. 2
• The bimanual console uses commercially available 6-DoF arms with a flange-mounted 2-DoF direct-drive wrist, continuous grasp sensing, and clutch input, at approximately US$5,500 in hardware cost.The work demonstrates one laparoscopic implementation; multi-interface and cross-vendor reconfiguration remain future work.Source: §I · pp. 1–2
• After 18 min of simulator familiarization, the operator used the console throughout the 32.5-min robotic portion without a master-console interruption; five operator-side metrics and hand/tip speeds were descriptively compared with 120 human in-vivo recordings.No abnormal clinical signs were recorded during the 7-day postoperative observation.Source: §I · p. 2

A. Surgical Master Interfaces

• Dedicated masters provide surgical functions as fixed assemblies, while stylus devices offer pose and a single trigger rather than directly exposed wrist articulation and separately sensed grip.Affordable add-on grippers and pedals still leave differences in hand interface and motion range relative to dedicated consoles.Source: A. Surgical Master Interfaces · p. 2

B. Kinematically Matched Robot-Like Leaders

• Matched or equivalent leader–follower kinematics support intuitive manipulation, but the trocar-constrained follower and general-purpose arm have no corresponding wrist or jaw joints.The system therefore exports tool pose, articulation, grasp, and clutch rather than mirrored joint coordinates.Source: B. Kinematically Matched Robot-Like… · p. 2

C. RCM-Constrained Surgical Teleoperation

• The pre-existing 1-kHz follower-side RCM controller is unchanged; the contribution lies upstream in the active master, re-anchoring, and endpoint-decoupled operator state.Prior approaches combine follower-side RCM enforcement with master-side haptic or hierarchical control.Source: C. RCM-Constrained Surgical… · p. 2

D. Modular Teleoperation and Surgical Research Platforms

• Where TELESIM, CRTK, and related platforms standardize robot-side access, this approach makes the operator-side message invariant so hand-interface and follower changes need not propagate across the stack.The supplied table passage identifies Table I as positioning representative teleoperation approaches but gives no cell values.Source: D. Modular Teleoperation and Surgical… · pp. 2–3, Table I · p. 3

A. System Overview

• The distal adapter is mechanically separable from the base arm; adaptation to another leader is intended to require a robot-specific flange and updated kinematic and payload models.Cross-vendor adaptation is not evaluated.Source: A. System Overview · p. 3

B. Robot-Arm-Based Active Master

• The base arms run under host-computer control at 200 Hz, with measured joint positions and velocities used to generate torque commands.The stock end-effector is replaced by the custom distal module, so the nominal distal-link inertial model does not exactly represent the assembled leader.Source: B. Robot-Arm-Based Active Master · p. 3, Table III · p. 3
• Joint-wise gravity scaling, residual trim, and host-applied viscous stabilization compensate for model and drive-side errors; left and right parameters are calibrated independently.The arms serve as gravity-compensated input devices, and no force feedback is rendered to the operator.Source: B. Robot-Arm-Based Active Master · p. 3

C. Distal Module Design and Continuous Grasp

• The module adds two orthogonal direct-drive wrist axes; grasp is a passively actuated, continuously sensed scalar mapped to endpoint jaw aperture, not an additional kinematic DoF.A magnetic rotary sensor measures the grasp shaft angle at 167 Hz; both wrist axes are exported, while the rigid in-vivo follower maps only corresponding twist to instrument roll.Source: C. Distal Module Design and Continuous… · p. 3

D. Direct-Drive Wrist Identification and Compensation

• The wrist drive reported 0.631±0.022 of commanded current; a 4-g load at a 50-mm lever arm yielded Kt = 0.0613 N m/A.The torque constant was estimated from bidirectional break-away averages before and after loading, and subsequent current results use the corrected scale.Source: D. Direct-Drive Wrist Identification… · p. 3
• The distal module has a flange-mounted 2-DoF articulated wrist and a sensor–magnet arrangement that samples grasp at 167 Hz.The module figure also shows passive thumb/index levers and the implemented operator hand interface.Source: Figure 3 · p. 4
• Across 13 tested roll/yaw postures, bidirectional break-away measurements separated friction and gravity currents, with friction exceeding gravity throughout the tested workspace.The figure also compares current-flow duty, estimated copper loss, and motor-case temperature for the earlier current-on and final motion-gated holding tests.Source: Figure 4 · p. 4, D. Direct-Drive Wrist Identification… · p. 4
• The final motion-gated controller applies posture-dependent gravity compensation and gates friction assistance on detected operator motion, allowing output to fall near zero at rest.This avoids applying fixed current continuously.Source: D. Direct-Drive Wrist Identification… · p. 5

E. Endpoint-Independent Interface and Mapping

• The common master state contains tool pose, wrist articulation qw, grasp g, and clutch state c.The same state is serialized as a 96-B ROS 2 message at 200 Hz over DDS for the physical path; Isaac Sim consumes it through the simulation endpoint.Source: E. Endpoint-Independent Interface and… · p. 5
• At clutch engagement, master and follower poses are stored as anchors for mapping translation.When released, the follower target tracks the measured tip so re-engagement re-establishes the offset at the instrument’s current pose.Source: E. Endpoint-Independent Interface and… · p. 5
• Physical mapping uses s = 1/3 and sends only incremental twist about the tool axis as instrument roll; the trocar RCM sets shaft pitch and yaw.Frame conversion, scaling, validity checks, and clutch anchoring remain in the endpoint adapter.Source: E. Endpoint-Independent Interface and… · p. 5

F. Simulation Environment

• The master-state interface drove an Isaac Sim scene with a rabbit gallbladder phantom and dual FR3 models constrained by fixed trocar ports.This setup supported familiarization and endpoint verification.Source: F. Simulation Environment · p. 5

A. Bench Hardware and Teleoperation Tests

• Bench tests compared current-flow duty, estimated copper loss, and motor-case temperature between an earlier current-on implementation and the final motion-gated controller.The test program also checked continuous grasp against the simulated jaw command and measured tip tracking during line, circular, and free 3-D motions.Source: A. Bench Hardware and Teleoperation… · p. 5
• Tip-tracking error was measured against the FR3-base-frame tip position after best delay alignment, using samples moving faster than 2 mm/s.Four clutch trials measured master repositioning and tool motion across release and re-engagement.Source: A. Bench Hardware and Teleoperation… · p. 5

B. Communication and Integration Timing

• Communication testing transmitted the 96-B ROS 2 master-state message at 200 Hz for 30 s over dedicated Ethernet and measured packet delivery and sender-clock RTT separately from robot motion.Master-state publication-to-FR3 target-update dead time included transport, node scheduling, mapping, and target publication, but excluded later controller, actuator, and mechanical response.Source: B. Communication and Integration Timing · p. 5

C. In-Vivo Protocol

• One veterinarian received 18 min of Isaac Sim familiarization before a single porcine cholecystectomy, with robotic work focused on gallbladder manipulation and dissection.The robotic interval lasted 32.5 min; a 3.7-min patient-side feedback stall was excluded, leaving 28.4 min for motion analysis.Source: C. In-Vivo Protocol · p. 5
• Clinical context came from Versius telemetry covering 120 recordings with required channels from 99 cholecystectomy procedures, pooled descriptively because some recordings shared procedures.Five operator-side metrics were defined consistently across the telemetry and porcine session; hand-speed and commanded-tip-speed distributions were analyzed separately.Source: C. In-Vivo Protocol · pp. 5–6
• The Versius data were used only as a descriptive reference, not as a matched control or equivalence benchmark.Re-anchoring continuity was evaluated across 46 valid in-vivo releases outside the interruption interval; the animal was monitored daily for 7 days after surgery.Source: C. In-Vivo Protocol · p. 6
• In-vivo tip tracking used engaged samples at 100 Hz, with roll error delay-aligned when commanded roll rate exceeded 5°/s.Roll error was the shaft-axis angle between the measured tool axis and the controller’s parallel-transported roll reference.Source: C. In-Vivo Protocol · p. 6

A. Distal-Module Hardware Characterization

• Friction exceeded gravity at all 13 tested wrist postures, and the grasp sensor resolved one 12-bit count, or 0.45% of full right-hand opening.The held lever showed no count change in 99.8% of 1-s windows when within 2% of full opening.Source: A. Distal-Module Hardware… · p. 6
• 89% to 11% current-flow duty and 2.7 to 0.3 W estimated copper loss followed the change to the final holding strategy.The earlier current-on baseline rose from 37°C to 52°C within 2 min, while the final motion-gated holding test remained near 36°C over 8 min.Source: A. Distal-Module Hardware… · p. 6

B. Bench Teleoperation and Integration

• 0.31–0.66 mm mean delay-aligned tip-tracking errors followed removal of the 230–235 ms delay, with p95 below 1 mm for circular and free-space motion and 2.07 mm for reversing line motion.The measured delay came from the pre-existing follower controller; the master-to-target bound excludes downstream controller and mechanical response.Source: B. Bench Teleoperation and Integration · p. 6
• At most 8 ms elapsed from master-state publication to FR3 target-command update, excluding patient-side controller and actuator response.The measured path included transport, scheduling, mapping, and target publication.Source: B. Bench Teleoperation and Integration · p. 6
• 7.5–14.1 cm of master repositioning produced 0.01–0.04 mm tool motion during clutch release and 1.11–3.34 mm immediately after re-engagement across four trials.The post-release motion occurred as the instrument followed hand motion already underway when the pedal was pressed.Source: B. Bench Teleoperation and Integration · p. 6

C. Porcine In-Vivo Evaluation and Clinical Context

• No interruption originated from the master console; across 46 valid re-anchorings, median instrument motion during release was 0.10 mm and median target-command discontinuity at re-engagement was 11 µm.The right hand was repositioned by a median 56.5 mm, and 42 of 46 right-side releases stayed below 1 mm instrument motion.Source: C. Porcine In-Vivo Evaluation and… · p. 6, Figure 7 · p. 7
• All five operator-side metrics fell within the range spanned by 120 Versius recordings, but two were outside the interquartile range: manipulating-time share was −0.92 IQR and handle angular speed was +1.65 IQR.The comparison is descriptive, not an equivalence test; hand speed overlapped the pooled clinical curve while commanded tip speed was shifted slower.Source: C. Porcine In-Vivo Evaluation and… · p. 6, Figure 7 · p. 7
• Removing a 230-ms delay reduced median in-vivo tip error from 1.28 to 0.28 mm on the right and from 0.66 to 0.13 mm on the left; the animal completed 7 days without abnormal clinical signs recorded.The residual increased with tip speed; right-hand roll had a median residual of 0.32° and p95 of 6.7°.Source: C. Porcine In-Vivo Evaluation and… · p. 6

VI. DISCUSSION AND LIMITATIONS

• The operator-side state drives either simulation or the RCM-constrained dual-FR3 platform through endpoint-specific adapters, while endpoint-specific safety and RCM control remain local.The master-to-target update bound is therefore limited to the master-side path, not the patient-side response.Source: §VI · p. 7
• The rigid laparoscopic follower has no distal articulation, so only instrument roll was mapped in vivo and orientation fidelity is reported for roll only.The master wrist has two actively articulated degrees of freedom.Source: §VI · p. 7
• The in-vivo evidence is one porcine case with one operator, compared descriptively with human procedures on a different platform; overlap does not establish equivalence or clinical readiness.Formal usability conclusions require multiple procedures, operators, and workload measures.Source: §VI · p. 8

• The master combines general-purpose robot arms with a separable adapter providing wrist articulation, continuous grasp input, and clutch-based workspace management.The operator-side state supports both simulation and an RCM-constrained physical platform.Source: §VII · p. 8
• Hardware characterization informed a low-duty hand-guided controller, and an endpoint-independent master state supported simulation and a physical RCM-constrained platform.Bench and porcine evaluation showed low-residual command transfer and stable re-anchoring.Source: §VII · p. 8
• Common operator-side metrics and hand/tip-speed distributions were placed in descriptive context against human clinical Versius telemetry.The authors conclude that the platform provides an accessible and modifiable foundation for future laparoscopic teleoperation research.Source: §VII · p. 8