cs.ROMar 12, 2026

Dynamic Model Identification and Gravity Compensation for the dVRK-Si Patient Side Manipulator

Authors: Haoying Zhou, Hao Yang, Brendan Burkhart, Anton Deguet, Zhonghao Zhang, Loris Fichera, Gregory S. Fischer, Jie Ying Wu, 

Organizations: Department of Robotics Engineering, Worcester Polytechnic Institute, MA 01609, USA. · Laboratory for Computational Sensing and Robotics, Johns Hopkins University, MD 21218, USA. · Department of Computer Science, Vanderbilt University, TN 37212, USA. · Department of Computer Science, Johns Hopkins University, MD 21218, USA.

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

The da Vinci Research Kit (dVRK) has been widely used for research in robot-assisted surgery over the past decade, and its latest generation, the dVRK-Si was made publicly available in 2025. This manuscript presents the first complete kinematic and dynamic modeling framework for the dVRK-Si Patient-Side Manipulator (PSM), whose redesigned mechanism differs substantially from that of the first-generation dVRK and therefore requires a new formulation to accurately capture gravitational effects and improve overall control performance. We first develop a modified DH kinematic model that represents the dVRK-Si PSM's redesigned mechanism. The equations of motion are then derived using the Euler-Lagrange formulation, and the inverse dynamics are expressed in a linear-in-parameters regressor form. The dynamic parameters are identified from data collected along a periodic excitation trajectory optimized for numerical conditioning and estimated by convex optimization subject to the physical feasibility constraints. The identified model is integrated into the dVRK control stack to enable real-time gravity compensation and computed-torque feedforward control. Experiments on a physical dVRK-Si demonstrate that the gravity compensation reduces steady-state joint errors by up to 84% and decreases end-effector tip drift during static holds from 4.2 mm to 0.7 mm. Computed-torque feedforward control further improves transient and trajectory-tracking accuracy. During sinusoidal trajectory tracking, computed-torque feedforward reduces position errors by 35% versus gravity-only feedforward and by 40% versus PID-only. The proposed pipeline could support reliable control, high-fidelity simulation, and learning-based automation on the dVRK-Si.

Figures & tables

Fig. 1: Patient Side Manipulator (PSM): dVRK Classic (left), dVRK-Si (right).
Fig. 2: An overview of the workflow for the proposed dVRK-Si PSM modeling, dynamic parameter identification, and deployment. Free-space dynamic model identification proceeds from symbolic robot modeling to optimized periodic excitation-trajectory generation, hardware execution, and physically consistent parameter estimation via convex optimization to obtain the full dynamic model. The identified model is then deployed through three control implementations: (1) a simplified statics-only gravity-compensation variant fitted from static torque measurements (neglecting friction and spring effects) and integrated into the dVRK software, (2) real-time open-loop gravity compensation implementation, and (3) real-time computed-torque feedforward, where joint accelerations are obtained by numerical differentiation and inverse-dynamics torques are added as a feedforward term to a baseline PID loop.
Fig. 3: Overview of the joint coordinates provided by dVRK software (n=7), extended for dynamics estimation (n=13), and for complete model (n=20).
Fig. 4: Modeling of the motion of the PSM gripper.
Fig. 5: The planar view of the frame definition for the dVRK-Si PSM. The units of the link lengths are mm. All frames are left-hand frames.
iiRefai−1a_{i-1}αi−1\alpha_{i-1}did_{i}θi\theta_{i}δLi\delta_{Li}ImiI_{mi}FiF_{i}KsiK_{si}
100π2\frac{\pi}{2}0q1+π2q_{1}+\frac{\pi}{2}✓\checkmark×\times✓\checkmark×\times
1’1−l1H-l_{1H}−π2-\frac{\pi}{2}0π2\frac{\pi}{2}×\times×\times×\times×\times
21’l1Ll_{1L}00q2−π2q_{2}-\frac{\pi}{2}✓\checkmark×\times✓\checkmark×\times
2’2l2L0l_{2L0}π2\frac{\pi}{2}l2H0l_{2H0}0×\times×\times×\times×\times
2”2’0−π2-\frac{\pi}{2}0−q2+π2-q_{2}+\frac{\pi}{2}✓\checkmark×\times✓\checkmark×\times
2”’2”l2L1l_{2L1}π2\frac{\pi}{2}l2H1l_{2H1}0×\times×\times×\times×\times
TABLE I: dVRK-Si PSM Model Description Parameters
Fig. 6: Experimental setup: dVRK-Si PSM is mounted at a zero-degree inclination relative to the table.
PlacementJ1J2J3J4J5J6J7
0 ∘ Horizontal (JHU)7.4310.5721.9136.2845.2945.4987.56
0 ∘ Horizontal (VU)10.727.6014.5548.9031.9937.8987.04
45 ∘ Downward (VU)9.457.9220.5162.1440.0164.4968.47
TABLE II: NRMSE Results for All Joints under 0 ∘ Horizontal Placement and 45 ∘ Downward-Tilted Surgical Scene Placement
Fig. 7: Measured and model-predicted joint torques/forces over a representative segment of an excitation trajectory.
ItemPose IdxPD Pos ErrGC Pos ErrPD τ\mathbf{\tau}GC τ\mathbf{\tau}Non- Drift LB τ\mathbf{\tau}Non- Drift UB τ\mathbf{\tau}
Joint1 (deg or N ⋅\cdot m)10.140.04-1.42-1.07-0.11-1.76
20.060.040.601.250.432.17
30.850.168.877.676.069.75
40.500.11-5.25-6.39-4.66-7.99
50.270.05-2.81-3.61-2.48-4.51
Joint2 (deg or N ⋅\cdot m)10.500.135.294.322.985.75
TABLE III: Drift Test Results
Fig. 8: Sinusoidal trajectory tracking using PID controller only (top), PID controller with dVRK software gravity compensation (middle), and PID controller with computed torque feedforward (bottom). Left panels compare desired and measured joint positions and report the position tracking error qerr=qdes−qmeas\bm{q}_{err}=\bm{q}_{des}-\bm{q}_{meas} . Right panels show the corresponding torque/force, including the desired joint efforts, measured joint efforts, and the errors between the measured and desired joint efforts τerr=τdes−τmeas\bm{\tau}_{err}=\bm{\tau}_{des}-\bm{\tau}_{meas} over a 10 s trial.
ItemJoint IdxMeanSTDRMSEMax
Joint Position Errors (deg or mm)
PID with Torque Feedforward10.460.240.520.94
20.440.220.491.07
30.780.400.881.57
PID with dVRK Software Gravity Compensation10.680.300.741.13
20.660.300.731.14
TABLE IV: Tracking Error Statistics
Fig. 9: Residual PID feedback joint torque/force comparison; contributions due to gravity compensation and computed-torque feedforward are excluded.
Fig. 10: Comparison of two gravity-compensation implementations (from Section III-I1 and Section III-J ) and the full computed-torque term over a sinusoidal trajectory.
More in this topic

Robot Dynamics Identification

Robot dynamics identification estimates the masses, inertias, friction, actuator behavior, and other forces that govern robot motion.

See all 58 papers

Explore similar work

CardsList
May 15, 2026

A Reproducible and Physically Feasible Dynamic Parameter Identification Framework for a Low-Cost Robot Arm

This paper presents a reproducible and physically feasible dynamic parameter identification framework for CRANE-X7, a low-cost robot arm driven by modular smart actuators. To improve practical identifiability, products of inertia are removed according to approximate link symmetry, reducing the rigid-body model from 65 to 39 base parameters. Identification motions are hand-designed from structured single-joint and adjacent-joint primitives under practical joint-range limits. The proposed pipeline combines preprocessing, inverse-dynamics-regressor-based ordinary least squares (OLS), conditional semidefinite-programming (SDP) projection for feasibility recovery, and closed-loop input error (CLIE) refinement. Candidate solutions from 40 structured trajectories are analyzed in a common principal component analysis (PCA) space to select a statistically central representative model. Because statistical centrality alone does not ensure physical acceptability, the selected model is finally screened by an all-pose positive-definiteness audit of the inertia matrix and, when necessary, corrected by a localized post-CLIE SDP rescue step. Experiments show that the parameter cloud becomes progressively more concentrated from OLS to SDP and CLIE, while the final accepted model preserves high predictive accuracy on held-out validation motions. These results demonstrate a practical route to statistically coherent and physically feasible dynamic models for low-cost robot platforms.

Junji Oaki, Koki Yamane, Koki Inami +1

Institute of Systems and Information Engineering, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8573, Japan

Robot Dynamics IdentificationParameter IdentifiabilityNonlinear System IdentificationSystem IdentificationRobot System Identification
AskPDF
Mar 15, 2026

Coordinate-Independent Robot Model Identification

Robot model identification is commonly performed by least-squares regression on inverse dynamics, but existing formulations measure residuals directly in coordinate force space and therefore depend on the chosen coordinate chart, units, and scaling. This paper proposes a coordinate-independent identification method that weights inverse-dynamics residuals by the dual metric induced by the system Riemannian metric. Using the force--velocity vector--covector duality, the dual metric provides a physically meaningful normalization of generalized forces, pulling coordinate residuals back into the ambient mechanical space and eliminating coordinate-induced bias. The resulting objective remains convex through an affine-metric and Schur-complement reformulation, and is compatible with physical-consistency constraints and geometric regularization. Experiments on an inertia-dominated Crazyflie--pendulum system and a drag-dominated LandSalp robot show improved identification accuracy, especially on shape coordinates, on both downsampled and full datasets.

Yanhao Yang, Ross L. Hatton

Collaborative Robotics and Intelligent Systems (CoRIS) Institute at Oregon State University, Corvallis, OR USA

Robot Dynamics IdentificationRobot System IdentificationSystem IdentificationInverse Dynamics
AskPDF
Mar 16, 2026

A Methodology for Dynamic Parameters Identification of 3-DOF Parallel Robots in Terms of Relevant Parameters

The identification of dynamic parameters in mechanical systems is important for improving model-based control as well as for performing realistic dynamic simulations. Generally, when identification techniques are applied only a subset of so-called base parameters can be identified. More even, some of these parameters cannot be identified properly given that they have a small contribution to the robot dynamics and hence in the presence of noise in measurements and discrepancy in modeling, their quality of being identifiable decreases. For this reason, a strategy for dynamic parameter identification of fully parallel robots in terms of a subset called relevant parameters is put forward. The objective of the proposed methodology is to start from a full dynamic model, then simplification concerning the geometry of each link and, the symmetry due to legs of fully parallel robots, are carried out. After that, the identification is done by Weighted Least Squares. Then, with statistical considerations the model is reduced until the physical feasibility conditions are met. The application of the propose strategy has been experimentally tested on two difierent configurations of actual 3-DOF parallel robots. The response of the inverse and forward dynamics of the identified models agrees with experiments. In order to evaluate the forward dynamics response, an approach for obtaining the forward dynamics in terms of the relevant parameters is also proposed.

Miguel Díaz-Rodríguez, Vicente Mata, Angel Valera +1

Departamento de Mecánica y Materiales, Universidad Politécnica de Valencia, 46022, Camino de Vera, Valencia, Spain · Departamento de Ingeniería de Sistemas y Automática, Universidad Politécnica de Valencia, 46022, Camino de Vera, Valencia, Spain · Departamento de Física Aplicada, Universidad Politécnica de Valencia, 46022, Camino de Vera, Valencia, Spain

Robot Dynamics IdentificationRobot System IdentificationRoboticsParameter Identifiability
AskPDF