3D force control for robotic-assisted beating heart surgery based on viscoelastic tissue model

International audience

Force controller design and evaluation for robot-assisted die and mould polishing

SUMMARY Linear and non-linear viscoelastic (VE) models such as the standard linear solid (SLS) and the generalized SLS (GSLS) are broadly used to represent the anelasticity of materials and Earth's media. However, although the VE approach is often satisfactory for any given observation, the inferred physical causes of anelasticity may be significantly misrepresented by this paradigm, and its predictions may be wrong or inaccurate in other cases. This problem is particularly important in heterogeneous media, including most cases of interest for seismology. For example, in homogenous media, VE and mechanics-based models predict identical quality-factor Q(f) and phase velocity c(f) spectra, but in heterogenous media, these models yield different time-stepping equations and interactions with material–property boundaries. The commonly used VE algorithms for modelling seismic waves rely on postulated convolutional integrals in time, whereas physically, models of rock rheologies should still be based on spatial interactions. To understand how VE models relate to mechanics, it is instructive to consider which physical properties of the medium are constrained reliably and which of them remain unconstrained by a pair of Q(f) and c(f) spectra, that is by VE properties. Despite its popular association with ‘attenuation,’ the peak value of Q−1(f) is actually a purely elastic property representing the existence of two (for SLS) or multiple (for GSLS) elastic moduli. These moduli are analogous to the drained and undrained moduli in poroelasticity or isothermal and adiabatic moduli in thermodynamics. By virtue of the Kramers–Krönig relations, the peak Q−1 is related to the total velocity dispersion, which is also caused by the difference between elastic moduli. By contrast, true anelasticity-related physical properties like viscosity are represented not by Q−1 values but by the frequencies of Q−1(f) peaks in the data. However, these frequencies also depend on multiple material properties that are not recognized or arbitrarily selected in the SLS and GSLS models. Inertial, body-force friction and the corresponding boundary effects are also ignored in VE models, which may again be improper for layered media. Thus, for physically accurate interpretation of laboratory experiments and numerical modelling of seismic waves, first-principle equations of mechanics should be used instead of VE models.

Constrained motion of a planar rigid-flexible manipulator is considered in this study. a dynamic model with zero tip deformation constraint of constrained rigid-flexible 608

We propose a novel autonomous robotic palpation framework for real-time elastic mapping during tissue exploration using a viscoelastic tissue model. The method combines force-based parameter estimation using a commercial force/torque sensor with an ergodic control strategy driven by a tailored Expected Information Density, which explicitly biases exploration toward diagnostically relevant regions by jointly considering model uncertainty, stiffness magnitude, and spatial gradients. An Extended Kalman Filter is employed to estimate viscoelastic model parameters online, while Gaussian Process Regression provides spatial modelling of the estimated elasticity, and a Heat Equation Driven Area Coverage controller enables adaptive, continuous trajectory planning. Simulations on synthetic stiffness maps demonstrate that the proposed approach achieves better reconstruction accuracy, enhanced segmentation capability, and improved robustness in detecting stiff inclusions compared to Bayesian Optimisation-based techniques. Experimental validation on a silicone phantom with embedded inclusions emulating pathological tissue regions further corroborates the potential of the method for autonomous tissue characterisation in diagnostic and screening applications.

This thesis aims to advance our understanding of ultrasound-driven microbubble dynamics, focusing on interactions with viscoelastic media and confined environments. By combining theoretical modeling and experimental observations, we explore microbubble resonance, scattering, and oscillation across different configurations, including bulk media and microchannels. The findings aim to improve the use of microbubbles in clinical applications, enhancing ultrasound imaging and therapeutic precision.<br/><br/>After introducing bubble dynamics and viscoelastic tissue models, Chapter 2 investigates how microbubble oscillations influence the rheological properties of the surrounding medium. Tissue behavior at high strain rates (106 s-1) - typical of microbubble activity - is not well understood. To address this, we examine polyacrylamide (PAM) hydrogels as tissue mimicking material, using microbubbles as probes. By analyzing ultra-high-speed imaging data and fitting them to coated bubble simulations, we extract the shear modulus and viscosity of soft and stiff hydrogels. Results show viscosity is highly strain-rate dependent, while the elastic modulus remains comparable to low-frequency values.<br/><br/>In Chapter 3, the focus shifts to how the viscoelastic medium’s shear modulus influences the behavior of phospholipid-coated microbubbles. Using a Rayleigh-Plesset-type model, we simulate the scattered pressure from bubbles and study the impact of medium elasticity on resonance frequency, harmonic generation, and subharmonic oscillations.<br/><br/>Chapter 4 builds on this by introducing a microbubble scattering spectroscopy technique to experimentally validate simulation findings. We correlate hydrogel stiffness with microbubble resonance frequency, quantify damping, and demonstrate how the gels affect microbubble stability.<br/><br/>Chapter 5 explores microbubbles in confined environments, such as small blood vessels, using silicone microchannels (15-100 µm). Ultra-high-speed imaging reveals a decrease in resonance frequency and increased damping as confinement increases. Optical data on radial oscillations is compared with theoretical predictions.<br/><br/>Together, these studies enhance our understanding of how microbubbles interact with soft and structured environments, providing a foundation for more accurate clinical models and improved therapeutic outcomes.

The classic approach to structure from motion entails a clear separation between motion estimation and structure estimation and between two-dimensional (2D) and three-dimensional (3D) information. For the recovery of the rigid transformation between different views only 2D image measurements are used. To have available enough information, most existing techniques are based on the intermediate computation of optical flow which, however, poses a problem at the locations of depth discontinuities. If we knew where depth discontinuities were, we could (using a multitude of approaches based on smoothness constraints) accurately estimate flow values for image patches corresponding to smooth scene patchess but to know the discontinuities requires solving the structure from motion problem first. This paper introduces a novel approach to structure from motion which addresses the processes of smoothing, 3D motion and structure estimation in a synergistic manner. It provides an algorithm for estimating the transformation between two views obtained by either a calibrated or uncalibrated camera. The results of the estimation are then utilized to perform a reconstruction of the scene from a short sequence of images. The technique is based on constraints on image derivatives which involve the 3D motion and shape of the scene, leading to a geometric and statistical estimation problem. The interaction between 3D motion and shape allows us to estimate the 3D motion while at the same time segmenting the scene. If we use a wrong 3D motion estimate to compute depth, we obtain a distorted version of the depth function. The distortion, however, is such that the worse the motion estimate, the more likely we are to obtain depth estimates that vary locally more than the correct ones. Since local variability of depth is due either to the existence of a discontinuity or to a wrong 3D motion estimate, being able to differentiate between these two cases provides the correct motion, which yields the “least varying” estimated depth as well as the image locations of scene discontinuities. We analyze the new constraints, show their relationship to the minimization of the epipolar constraint, and present experimental results using real image sequences that indicate the robustness of the method.

This paper proposes an intelligent massage control system by using multi-fingered robot hand with hybrid impedance and force control, which is able to create the movement and force of robot to behave as similar as the human’s massage. Therefore, the various massage points, such as the changing of hardness and softness of human skin muscle condition, can be appropriately controlled by using hybrid impedance and force control method. The impedance control is used to control the lateral position of massage on the human skin muscle. The perpendicular direction one, while force control is carried out for the vertical direction on human skin muscle. This paper also gives a modeling of human skin muscle through robot perception of impedance to decide the parameter of impedance controller. The control strategy using a hybrid impedance and force control to implement an adaptive control system is presented, when human condition is changed with soft and hard skin muscle. For the robot fingertip force, PI servo compensation is used together with impedance control. Based on using both impedance and PI servo compensation, the effectiveness of the massage position and force control is shown through actual massage motion experiments.

Each joint of a hydraulic-driven legged robot adopts a highly integrated hydraulic drive unit (HDU), which features a high power–weight ratio. However, most HDUs are throttling-valve-controlled cylinder systems, which exhibit high energy losses. By contrast, pump control systems offer a high efficiency. Nevertheless, their response ability is unsatisfactory. To fully utilize the advantages of pump and valve control systems, in this study, a new type of pump–valve compound drive system (PCDS) is designed, which can not only effectively reduce the energy loss, but can also ensure the response speed and response accuracy of the HDUs in robot joints to satisfy the performance requirements of robots. Herein, considering the force control requirements of energy conservation, high precision, and fast response of the robot joint HDU, a nonlinear mathematical model of the PCDS force control system is first introduced. In addition, pressure–flow nonlinearity, friction nonlinearity, load complexity and variability, and other factors affecting the system are considered, and a novel force control method based on quantitative feedback theory (QFT) and a disturbance torque observer (DTO) is designed, which is denoted as QFT–DTOC herein. This method improves the control accuracy and robustness of the force control system, reduces the effect of the disturbance torque on the control performance of the servo motor, and improves the overall force control performance of the system. Finally, experimental verification is performed using the PCDS performance test platform. The experimental results and quantitative data show that the QFT–DTOC proposed herein can significantly improve the force control performance of the PCDS. The relevant force control method can be used as a bottom-control method for the hydraulic servo system to provide a foundation for implementing the top-level trajectory planning of the robot.

SUMMARYThis paper proposes a detailed investigation on the new neural-based feed-forward PID direct force control (FNN-PID-DF) approach applied to a highly nonlinear 2-axes pneumatic artificial muscle (PAM) manipulator in order to ameliorate its force output performance. Founded on the novel inverse neural NARX model dynamically identified to learn well all nonlinear characteristics of the contact force dynamics of the 2-axes PAM-based manipulator, the novel proposed neural FNN-PID-DF force controller is innovatively implemented in order to directly force control the 2-axes PAM robot system used as a rehabilitation device subjected to internal systematic interactions and external contact force deviations. The performance of the experimental tests has proven the advantages and merits of the new force control method compared to the classical PID force control method. The new neural FNN-PID-DF force controller guides the wrist/hand of subject/patient to successfully generate the predefined desired force values.

Haptic devices need to render accurate force to a human operator in order to create a virtual environment from a remote real environment. Existing end-effector type haptic devices employ parallel mechanisms to achieve force interaction in multiple degrees-of-freedom (DoF). Those parallel mechanisms have non-constant Jacobian matrices that create difficulties in the actuator and controller design for accurate haptic force interaction. This paper presents a 3-DoF translational parallel manipulator (TPM) with a constant Jacobian matrix to facilitate the force and impedance controller design. Instead of using external force sensors, series elastic actuators (SEAs) are used for the proposed device to realize the force rendering. SEAs can render more accurate output force and stiffness while the inertia and friction problems of using a gearhead can be ignored. In this paper, the modeling, design, and control of the haptic device are established to show the merits of the proposed TPM. Initial force control experiments demonstrate the accuracy and speed of SEA-based force control. We expect that this novel haptic device can serve as an alternative for robotic teleoperations requiring accurate force sensing and control.

This paper presents a force control system design by model predictive control scheme. Recently, a lot of research has been carried out on the robots that take the place of human labor. One of the challenges of such robots is deformable object manipulation to working on food processing, surgery, etc. For handling deformable objects by a rigid robot manipulator is required an advanced reaction force control methodology. Such force control methods have to consider suitable force application without excess and deficiency forces for preventing to drop and destroy in contact with unknown objects. To realize this, it is required to directly consider constraints on control input, rate of input, velocity, etc. Model predictive control is known as one of the key techniques in control fields, and it has a potential for stable force control implementation by considering constraints. This paper attempts to design a new force control system by utilizing model predictive velocity control based on an augmented state-space model with a disturbance term. Finally, the performances of the proposed method via numerical simulation are shown.

In industries manufacturing metallic molds, various NC machine tools are used. We have already proposed a desktop NC machine tool with compliance control capability to automatically cope with the finishing process of LED lens molds. The NC machine tool has the ability to control the polishing force acting between an abrasive tool and a work piece. The force control method is called impedance model force control. The most effective gain is the desired damping of the impedance model. Ideally, the desired damping is calculated from the critical damping condition after considering the effective stiffness in the force control system. However, there is a problem in that the effective stiffness of the NC machine tool has undesirable nonlinearity. The nonlinearity has a bad influence on the force control stability. In this article, a fine tuning method of the desired damping is considered using neural networks. The neural networks acquire the nonlinearity of effective stiffness. The promise is evaluated through an experiment.

In this contribution two methods to increase the robustness of robot force control are discussed and experimental results on a real system are presented. One way to control end–effector forces of a robot is to use a force control method which manipulates the desired trajectory of a position controlled robot in a cascaded scheme. The benefit of this control scheme is that the end–effector is also able to follow position trajectories in the non–force–controlled directions. For such a direct parallel force/position control the stability depends on the force controller parameters and the contact stiffness, assuming a stable position control. Tuning these parameters in environments with varying contact stiffness is challenging and time consuming. To avoid this, additional feed–back torques are calculated to increase the dynamic of the force controller. The first method is based on the classical feed–forward control of a robotic system which is used in the feed–back loop of the force control. In the second method the acceleration term of the force control law is used to calculate a feed–back torque.

The paper proposes a feedback force and velocity control of an arm exoskeleton to assist user motion. The original published control so-called feedback hybrid force and position control was based on the force and position control and was designed to assist user motion. This original control was successful at providing assist for the user’s arm. This article presents an improved control scheme called the feedback force and velocity control. The proposed control is designed to regulate the velocities of joints of the exoskeleton and the feedback forces on links to assist user motion. The design and optimization of the feedback force and velocity control are realized by the Balancing Composite Motion Optimization (BCMO). The numerical method is realized in the paper to show that the proposed control is better than the original control in terms of less oscillation and fast response.