Source-linked AI summary
Realistic Haptic Rendering of Interacting Deformable Objects in Virtual Environments
Christian Duriez, Frédéric Dubois, Abderrahmane Kheddar, Claude Andriot
TL;DR
Existing haptic simulations often simplify contact and friction between deformable objects, limiting realistic rendering of multicontact phenomena. The paper uses Signorini and Coulomb laws with a Delassus-operator formulation and Gauss-Seidel iteration, combined with global corotational deformation. The implemented approach demonstrates stable and realistic 6D haptic feedback in a clipping task.
Problem
Existing haptic methods simplify contact and friction laws and do not adequately handle realistic multicontact phenomena among deformable objects.
Method
The paper combines Signorini contact, Coulomb friction, Delassus-operator linearization, and Gauss-Seidel iteration with global corotational deformation.
Results
Stable and realistic 6D haptic feedback was demonstrated through an implemented clipping-task simulation.
Takeaways & Limitations
The approach provides a fast implicit treatment of contact and friction that can be combined with existing fast deformable simulations.
Takeaways & Limitations
The paper’s detailed theory develops the contact solution for a pair of deformable objects, with extension to n-bodies demonstrated separately.
Abstract
from arXiv · showhide
A new computer haptics algorithm to be used in general interactive manipulations of deformable virtual objects is presented. In multimodal interactive simulations, haptic feedback computation often comes from contact forces. Subsequently, the fidelity of haptic rendering depends significantly on contact space modeling. Contact and friction laws between deformable models are often simplified in up to date methods. They do not allow a "realistic" rendering of the subtleties of contact space physical phenomena (such as slip and stick effects due to friction or mechanical coupling between contacts). In this paper, we use Signorini's contact law and Coulomb's friction law as a computer haptics basis. Real-time performance is made possible thanks to a linearization of the behavior in the contact space, formulated as the so-called Delassus operator, and iteratively solved by a Gauss-Seidel type algorithm. Dynamic deformation uses corotational global formulation to obtain the Delassus operator in which the mass and stiffness ratio are dissociated from the simulation time step. This last point is crucial to keep stable haptic feedback. This global approach has been packaged, implemented, and tested. Stable and realistic 6D haptic feedback is demonstrated through a clipping task experiment.
1 INTRODUCTION
Realistic haptic manipulation of multiple deformable objects requires contact and friction models that handle no-penetration, stick-slip behavior, and multicontact coupling. The paper develops a Gauss-Seidel-based approach integrated with real-time deformation computation to address these challenges.
- Motivation: Multiple deformable-object interactions remain inadequately solved because common methods rely on single-point interaction or simplified contact laws.These limitations hinder extension to complex multiobject scenes and realistic haptic rendering.
- Limitations of Existing Methods: Coulomb friction captures stick and slip states but is difficult to solve correctly for multicontact deformable objects.Previous approaches often compute friction explicitly from the previous timestep, while imposed-motion methods can produce sticky effects and drift during sticking.
- Motivation: Contact-space modeling strongly affects haptic rendering fidelity because haptic output primarily reflects potentially numerous, topologically complex contacts.The paper identifies realistic contact models as crucial to transparency and rendering fidelity.
- Limitations of Existing Methods: Penalty methods trade straightforward contact solving for dependence on arbitrary penalty factors, instability risks, or nonlinear stiff problems.Implicit penalty solutions require penalty factors much higher than the objects’ elasticity modulus.
- Limitations of Existing Methods: Imposed-motion methods are difficult to generalize across deformable objects and are problematic for friction, especially tangential motion with weak adherence.The paper notes that these methods are generally used in frictionless contexts and may cause sticky effects.
- Contribution: The paper targets several deformable objects with no-penetration constraints and friction, using an iterative Gauss-Seidel algorithm without approximating Coulomb’s law.The approach incorporates contact forces into the physically based simulation engine for haptic feedback.
3 MODELING CONTACT FROM THE SIGNORINI’S LAW
The paper applies Signorini’s contact law to real-time interactions between deformable bodies, representing contact forces and gaps within a finite-element formulation. Collision data are converted into a coupled linear complementarity problem for enforcing noninterpenetration.
- Contact model: Signorini’s law is used to resolve deformable-body contacts in a real-time multimodal haptic simulation.The paper identifies this as a first contribution for real-time resolution in computer haptics.
- Contact model: The law enforces noninterference through complementarity between the normal gap and contact stress.An active contact has zero gap and nonzero pressure, whereas a separated contact has positive gap and zero stress.
- Contact representation: Collision detection supplies contact points, barycentric positions, and optionally contact normals for each potential contact.The algorithm makes no specific collision-detection assumption, but its results depend on the accuracy of this input.
- Scope and assumption: Missed collision intersections can temporarily cause interpenetration, but rapid restoration may avoid instability when motions are not too fast.The paper assumes efficient collision detection for its multifrictional-contact solution.
- Finite-element formulation: Linear tetrahedral interpolation maps contact forces and node displacements between surface contacts and mesh nodes.Forces at contact points are distributed to the nodes of their supporting triangles using interpolation functions.
- Linear formulation: Free-motion gaps and mechanically induced displacements are combined into a linear relation between contact forces and constrained positions.The formulation stacks contact relations through matrices H1 and H2 and a mechanical compliance matrix.
4 COULOMB’S FRICTION LAW
The paper incorporates Coulomb friction to represent stick and slip in deformable-object contact. Because multicontact dynamic friction is nonlinear and coupled across contacts, it replaces friction-cone approximation with an iterative Gauss-Seidel-like resolution.
- Friction model: Coulomb’s law models stick and slip through tangential contact behavior, but is difficult to solve correctly for multiple contacts.Dynamic friction depends on the unknown tangential motion and force direction.
- Multicontact coupling: Friction and contact calculations are coupled because forces at one contact can modify other contacts through tangential and normal interactions.This coupling prevents direct expression of general multicontact dynamic friction as a linear complementarity problem.
- Cone approximation: Polyhedral friction-cone approximations preserve an LCP formulation but replace the actual cone with k-sided pyramids.The approach is described for implicit or semi-implicit multicontact simulations with dry friction.
- Cone approximation: With k = 8 pyramid faces, the frictional LCP is 10 times larger than the frictionless formulation.The global LCP has size m × (k + 2) for m contacts.
- Iterative resolution: The proposed Gauss-Seidel-like algorithm resolves multicontact friction without approximating Coulomb’s law.The paper presents this iterative strategy as an alternative to the asymptotically costly pyramid method.
5 GAUSS-SEIDEL-LIKE ALGORITHM
The paper adapts a Gauss–Seidel-like algorithm to resolve Signorini contact and Coulomb friction for multiple deformable-body contacts. Using the Delassus operator preserves mechanical coupling while improving precision and efficiency as contacts increase.
- Algorithm: The algorithm adapts Gauss–Seidel resolution to multicontact problems involving deformable bodies and friction.It resolves each contact block iteratively while freezing the contributions of other contacts.
- Precision and performance: The method preserves the nonlinear Coulomb friction cone instead of approximating it with k-sided pyramids.The comparison evaluates the exact friction-cone GS approach against an LCP formulation using pyramid approximations.
- Precision and performance: A gap below 5 percent between the GS and LCP solutions requires 16-sided pyramids in the test configuration.The tests use a small friction coefficient, μ = 0.1, to produce dynamic friction.
- Precision and performance: The GS approach becomes much more efficient than the pyramid-based approach as the number of instantaneous contacts increases.Its worst-case asymptotic complexity is O(m^2), compared with O(k × m^2) for k-sided pyramids.
- Mechanical coupling: The Delassus operator incorporates material, structural, and intercontact coupling into contact-force computation.This distinguishes the method from penalty approaches and supports contacts between models with different mechanical impedance.
- Scalability: The algorithm’s complexity depends on contact count and coupling rather than object shape or node count, while tolerance trades precision against computation time.Because the method is iterative, operation count also depends on convergence velocity and the stopping tolerance.
6 COROTATIONAL APPROACH
The corotational global approach combines rigid global motion with local deformation in contact space. By decoupling mass and stiffness in the Delassus operator, it supports stable haptic feedback while retaining compliant behavior and frictional contacts.
- Time-step constraints: Increasing stiffness without reducing the time step can violate the Shannon theorem for deformation frequencies, requiring time-step adaptation.When stiffness matrices are singular without Dirichlet conditions, the time step must also make mass and damping dominant over stiffness.
- Global corotational model: The corotational global model splits deformable-body motion into rigid global motion and local relative displacement.The two components are combined as compliance within the contact space.
- Global corotational model: The approach maps rigid motion into contact space through a Jacobian and combines it with local deformation to obtain linearized contact behavior.The Jacobian maps motion-space quantities and contact forces between generalized and contact coordinates.
- Stability: Mass and stiffness are decoupled in the Delassus operator, allowing stable haptic feedback and real-time simulation with an arbitrary time step.This addresses the dependence of conventional dynamic simulation time steps on the mass/stiffness ratio.
- Quasi-rigid behavior: As stiffness increases, the model’s behavior tends toward rigid-body motion while retaining physically plausible mechanical compliance.In the quasi-rigid limit, the deformable component helps resolve indeterminacy associated with Coulomb friction.
- Quasi-rigid behavior: The FEM-based compliance distinguishes this approach from methods that add small deformations to rigid objects to obtain unique and smoother solutions.The authors state that their compliance produces more plausible results through an FEM model.
7 FORCE FEEDBACK COMPUTATION AND STABLE HAPTIC FEEDBACK ISSUES
The paper couples haptic interfaces to deformable-object simulations and evaluates stable force feedback during snap-in tasks. The experiments illustrate how friction, deformation, and contact dynamics shape the rendered interaction and computational cost.
- Snap-In Task: The snap-in scenario uses a virtual 6-degree-of-freedom coupling between the operator, a deformable clip, and a rigid or deformable pipe.The task comprises pushing, unstable equilibrium, and final clipping phases.
- Snap-In Task: The pushing phase combines static-friction sticking, deformation resistance, and dynamic friction as the clip is forced toward the pipe.Resistance continues until the clip deformation reaches its maximum.
- Snap-In Task: During unstable equilibrium, reduced applied forces can cause the clip to return abruptly, especially when the pipe is rigid.Sufficient force carries the clip into the final clipping phase.
- Snap-In Task: With two deformable objects, static friction and clip deformation resist withdrawal before the clip exits the cylinders abruptly.The virtual coupling accumulates potential energy while the operator moves the interface.
- Stable Haptic Feedback: The corotational method supports a realistic 15g clip mass with a 3ms simulation time step, while implicit contact and friction resolutions handle rapid motion.The example uses polyethylene parameters of Young’s modulus E = 700MPa and Poisson coefficient = 0.35.
8 CONCLUSION AND FUTURE WORK
The paper presents an implicit, fast treatment of contact and friction for stable haptics with flexible virtual objects. It reports implementation and snap-in testing, while identifying performance optimization and other friction models as future work.
- Conclusion: The approach solves Signorini’s contact law and Coulomb’s friction law implicitly with a fast Gauss-Seidel-like algorithm.Contact treatment is separated from deformation behavior and can be combined with existing fast deformable simulations.
- Conclusion: A global corotational formulation is shown to be useful for real-time simulation and stable force-feedback rendering.The models and algorithms were implemented, packaged, tested, and evaluated in a virtual snap-in simulation.
- Future Work: Future work targets computation-process optimization, other dry-friction models, and application of the techniques to medical simulation.These directions define the stated extension scope of the current work.