Thanks to the low price, the use of a head-mounted device (HMD) equipped with a smartphone is currently a common setup for virtual reality (VR). Brain-computer interface (BCI) based on electroencephalography (EEG) is a promising technology to enrich the VR experience. However, the effect of using HMDs on the acquisition of EEG signals remains still unknown. In fact, the smartphone is placed close to the head where EEG sensors are located, thus the smartphone's electronics may perturb the acquisition of the EEG signal. In the present study, we compare the spectral properties of the EEG signal acquired on 12 subjects wearing a SamsungGear HMD equipped with a Samsung S6 smartphone turned on and off. Our study shows that there is no significant difference in the spectral properties of the EEG in these two experimental conditions. We conclude that a smartphone-based HMD is compatible with EEG technology. Some technical problems related to the concurrent use of a HMD and an EEG-based BCI are also discussed.
The trade-off between accuracy and computational performance is one of the central conflicts in real-time multibody simulations, much of which can be attributed to the method used to solve the constrained multibody equations. This paper examines four mixed linear complementarity problem (MLCP) algorithms when they are applied to physical problems involving frictional contact. We consider several different, and challenging, test cases such as grasping, stability of static models, closed loops, and long chains of bodies. The solver parameters are tuned for these simulations and the results are evaluated in terms of numerical accuracy and computational performance. The objective of this paper is to determine the accuracy properties of each solver, find the appropriate method for a defined task, and thus draw conclusions regarding the applicability of each method.
We introduce a practical iterative solver for mass-spring systems which can be trivially mapped to massively parallel architectures, in particular GPUs.We employ our solver for the interactive animation of virtual cloth and show that it is computationally fast, robust and scalable, making it suitable for real-time graphics applications. Under the assumption that the input system is represented by a quadrangular network of masses connected by springs, we first partition the particles into two independent sets. Then, during the animation, the dynamics of all the particles belonging to each set is computed in parallel. This enables a full Gauss-Seidel iteration in just two parallel steps, leading to an approximated solution of large mass-spring systems in a few milliseconds. We use our solver to accelerate the solution of the popular Projective Dynamics framework, and compare it with other common iterative solvers in the current literature.
Damping is an important ingredient in physics-based simulation of deformable objects. Recent work introduced new fast simulation methods such as Position Based Dynamics and Projective Dynamics. Explicit velocity damping methods currently used in conjunction with Position Based Dynamics or Projective Dynamics are simple and fast, but have some limitations. They may damp global motion or non-physically transport velocities throughout the simulated object. More advanced damping models do not have these limitations, but are slow to evaluate, defeating the benefits of fast solvers such as Projective Dynamics. We present a new type of damping model specifically designed for Projective Dynamics, which provides the quality of advanced damping models while adding only minimal computing overhead. The key idea is to define damping forces using Projective Dynamics' Laplacian matrix. In a number of simulation examples we show that this damping model works very well in practice. When used with a modified Projective Dynamics solver that uses a non-dissipative implicit midpoint integrator, our damping method provides fully user-controllable damping, allowing the user to quickly produce visually pleasing and vivid animations.
We propose a novel method to predict pressure values at boundary particles in incompressible divergence-free SPH simulations (DFSPH). Our approach employs Moving Least Squares (MLS) to predict the pressure at boundary particles. Therefore, MLS computes hyperplanes that approximate the pressure field at the interface between fluid and boundary particles. We compare this approach with two previous techniques. One previous technique mirrors the pressure from fluid to boundary particles. The other one extrapolates the pressure from fluid to boundary particles, but uses a gradient that is computed with Smoothed Particle Hydrodynamics (SPH). We motivate that gradient-based extrapolation is more accurate than mirroring. We further motivate that our proposed MLS gradient is less error prone than the SPH gradient at the boundary. In our experiments, we indicate artifacts in previous approaches. We show that these artifacts are significantly reduced with our approach resulting in simulation steps that can be twice as large compared to previous methods. We further present challenging and complex scenarios to illustrate the capabilities of the proposed boundary handling. CCS Concepts •Computing methodologies → Physical simulation; Massively parallel and high-performance simulations;
This paper presents a novel software framework for quantitative validation of physically-based deformable models in computer graphics. In the majority of previous studies, validation is qualitative (through visual plausibility), which is necessarily user subjective. The proposed framework facilitates construction of a single scalar output that quantifies the agreement between the complete time histories of test and reference models. Different models and comparison metrics can be easily included within the general framework. The framework is shown to yield a high accuracy score for a simplified model that can be analytically derived from the reference model, indicating that the framework is reliable. A lower score results when evaluating a more approximate, yet still visually plausible, model, demonstrating the objective sensitivity of the framework. The software framework can thus provide an objective measure of accuracy and a standardised way to quantitatively compare the accuracy of one method against another, whilst also supplying a quantitative rationale for trading accuracy and performance.
We propose an approach for real-time shallow water simulation, building upon the virtual pipes model with multi-layered heightmaps. Our approach introduces the use of extended pipes which resolve flow through fully-flooded passages, which is not possible using current multi-layered techniques. We extend the virtual pipe method with a physically-based viscosity model that is both fast and stable. Our viscosity model is integrated implicitly without the expense of solving a large linear system. The liquid is rendered as a triangular mesh surface built from a heightmap. We propose a novel surface optimization approach that prevents interpenetrations of the liquid surface with the underlying terrain geometry. To improve the realism of small-scale scenarios, we present a meniscus shading approach that adjusts the liquid surface normals based on a distance field. Our approach runs in real time on various scenarios of roughly 10 x 10 cm at a resolution of 0.5 mm, with up to five layers. CCS Concepts •Computing methodologies → Physical simulation;
Cumulus clouds are ubiquitous in outdoor naturalistic scenes. However, realistic and scalable cloud simulation continues to be a challenge in computer graphics. Computational fluid dynamics models (CFD) are computationally intensive and hence impractical for real time methods. Eulerian models [Harris et al. 2003] on the other hand require modeling even the transparent volume, placing serious restrictions on the domain size. More importantly, both these methods do not offer convenient handles on initial, border and atmospheric conditions which play a key role in the natural evolution of the cloud phenomena. We introduce Lagrangian framework for the cloud physics and rendering at landscape size. The use of Lagrangian formalism, aerology physics, and separation of environment from the dynamics, brings more robustness, offers meaningful high-level handles for user control - environment profile and evolution, cloud base and top.
In this paper we present a new position based approach for simulating rigid and flexible bodies with two-way coupling. This is achieved by expressing all the dynamics as constraints and running them in the same solver. Our main contribution is an accurate contact and Coulomb friction model based on a fixed point iteration of a cone complementarity problem. We formulate the problem as a nonlinear convex minimization at position level and solve it using a new accelerated form of projected Jacobi. We add elasticity to the constraints by means of regularization and show how to add more damping in a credible manner. We also use this viscoelastic model to build an accurate position-based finite element solver for soft bodies. The novelty of this solver is that it is no longer an approximation and it is based directly on the elasticity theory of continuous media.
Physically-based simulation of heterogeneous objects remains computationally-demanding for many applications, especially when involving haptic interaction with virtual environments. In this paper, we introduce a novel multiresolution approach for haptic interaction with heterogeneous deformable objects. Our method called "Elasticity-based Clustering" is based on the clustering and aggregation of elasticity inside an object, in order to create large homogeneous volumes preserving important features of the initial distribution. The design of such large and homogeneous volumes improves the attribution of elasticity to the elements of the coarser geometry. We could successfully implement and test our approach within a complete and real-time haptic interaction pipeline compatible with consumer-grade haptic devices. We evaluated the performance of our approach on a large set of elasticity configurations using a perception-based quality criterion. Our results show that for 90% of studied cases our method can achieve a 6 times speedup in the simulation time with no theoretical perceptual difference.
Inspired by kernel methods that have been used extensively in achieving efficient facial animation retargeting, this paper presents a solution to retargeting facial animation in virtual character's face model based on the kernel projection of latent structure (KPLS) regression between semantically similar facial expressions. Specifically, a given number of corresponding semantically similar facial expressions are projected into the latent space. By using the Nonlinear Iterative Partial Least Square method, decomposition of the latent variables is achieved. Finally, the KPLS is achieved by solving a kernalized version of the eigenvalue problem. By evaluating our methodology with other kernel-based solutions, the efficiency of the presented methodology in transferring facial animation to face models with different morphological variations is demonstrated.
The paper shows that the SPH boundary handling of Akinci et al. [AIA* 12] suffers from perceivable issues in planar regions due to deviations in the computed boundary normals and due to erroneous oscillations in the distance computation of fluid particles to the boundary. In order to resolve these issues, we propose a novel boundary handling that combines the SPH concept with Moving Least Squares. The proposed technique significantly improves the distance and normal computations in planar boundary regions, while its computational complexity is similar to Akinci's approach. We embed the proposed boundary handling into Implicit Incompressible SPH in a hybrid setting where it is applied at planar boundaries, while Akinci's technique is still being used for boundaries with complex shapes. Various benefits of the improved boundary handling are illustrated, in particular a reduced particle leakage and a reduced artificial boundary friction.
We address the task of computing solutions for a separating fluid-solid wall boundary condition model. We present an embarrassingly parallel, easy to implement, fluid LCP solver. We are able to use greater domain sizes than previous works have shown, due to our new solver. The solver exploits matrix-vector products as computational building blocks. We block the matrix-vector products in a way that allows us to evaluate the products, without having to assemble the full systems. Any iterative sub-solver can be used. Our work shows speedup factors ranging up to 500 for larger grid sizes.
Objective: Motivated by the limitations of being unable to provide feedback and adequately assess technical skills whilst training unsighted physical examinations, such as Digital Rectal Examinations (DRE), we present a see-through visualisation system that can be used with benchtop models widely available in medical schools. Methods: We use position and pressure sensors located on the examining finger and have implemented a Virtual Reality (VR) simulation learning tool consisting of registered 3D models of the benchtop, augmented with relevant surrounding pelvic anatomy. The proposed system was evaluated with six medical students and eleven consultants. Results: The system is stable, runs in real time, uses unobtrusive sensor coils and pads, is able to capture data from sensors at 40Hz and adequately translates and rotates the position of the examining finger aligned to the 3D models of the benchtop and surrounding anatomy. Both medical students and consultants recognised the educational value of being able to see-through and visualise surrounding relevant anatomy. Although novices are reported to be the group that could benefit the most from our system, it is crucial not to be over reliant on visual cues for too long and to develop a strategy for the adequate use of the see-through system. Conclusions: The proposed VR simulation system is intended to improve the experience of novices learning unsighted examinations by providing real-time feedback and visualisation, allowing trainees to reflect on their performance and permitting more adequate assessment of technical skills.
This paper presents a Differential Evolutionary (DE) algorithm for solving multi-objective kinematic problems (e.g., end-effector locations, centre-of-mass and comfort factors). Inverse kinematic problems in the context of character animation systems are one of the most challenging and important conundrums. The problems depend upon multiple geometric factors in addition to cosmetic and physical aspects. Further complications stem from the fact that there may be non or an infinite number of solutions to the problem (especially for highly redundant manipulator structures, such as, articulated characters). What is more, the problem is global and tightly coupled so small changes to individual link's impacts the overall solution. Our method focuses on generating approximate solutions for a range of inverse kinematic problems (for instance, positions, orientations and physical factors, like overall centre-of-mass location) using a Differential Evolutionary algorithm. The algorithm is flexible enough that it can be applied to a range of open ended problems including highly non-linear discontinuous systems with prioritisation. Importantly, evolutionary algorithms are typically renowned for taking considerable time to find a solution. We help reduce this burden by modifying the algorithm to run on a massively parallel architecture (like the GPU) using a CUDA-based framework. The computational model is evaluated using a variety of test cases to demonstrate the techniques viability (speed and ability to solve multi-objective problems). The modified parallel evolutionary solution helps reduce execution times compared to the serial DE, while also obtaining a solution within a specified margin of error (<1%).
Computing the forces acting from a surrounding air phase onto an SPH free-surface fluid is challenging. For full multiphase simulations the computational overhead is significant and stability issues due to the high density ratio may arise. In contrast, the air-fluid interactions can be approximated efficiently by employing a drag equation. Here, for plausible effects, the parameterization is important but challenging. We present an approach to calculate the parameters of the used drag equation in a physically motivated way. We approximate the deformation and occlusion of particles to determine their drag coefficient and exposed surface area. The resulting effects are validated by comparing them to the results of a multiphase SPH simulation. We further show the practicality of our approach by combining it with different types of SPH solvers and by simulating multiple, complex scenes.
In this paper, we present a novel approach for a tighter integration of 3D modeling and physically-based simulation. Instead of modeling 3D objects as surface models, we use a volumetric subdivision representation. Volumetric modeling operations allow designing 3D objects in similar ways as with surface-based modeling tools. Encoding the volumetric information already in the design mesh drastically simplifies and speeds up the mesh generation process for simulation. The transition between design, simulation and back to design is consistent and computationally cheap. Since the subdivision and mesh generation can be expressed as a precomputable matrix-vector multiplication, iteration times can be greatly reduced compared to common modeling and simulation setups. Therefore, this approach is especially well suited for early-stage modeling or optimization use cases, where many geometric changes are made in a short time and their physical effect on the model has to be evaluated frequently. To test our approach, we created, simulated and adapted several 3D models. Additionally, we measured and evaluated the timings for generating and applying the matrices for different subdivision levels. For comparison, we also measured the tetrahedral meshing functionality offered by CGAL for similar numbers of elements. For changing topology, our implicit meshing approach proves to be up to 70 times faster than creating the tetrahedral mesh only based on the outer surface. Without changing the topology and by precomputing the matrices, we achieve a speed-up of up to 2800.
We present the first mesh-based surface tracker that runs entirely on the GPU. The surface tracker is both completely grid-free and fast which makes it suitable for the use in a large, unbounded domain. The key idea for handling topological changes is to detect and delete overlapping triangles as well as triangles that lie inside the volume. The holes are then joined or closed in a robust and efficient manner. Good mesh quality is maintained by a mesh improvement algorithm. In this paper we describe how all these steps can be parallelized to run effi-ciently on a GPU. The surface tracker is guaranteed to produce a manifold mesh without boundary. Our results show the quality and efficiency of the method in both Eulerian and Lagrangian liquid simulations. Our parallel implementation runs more than an order of magnitude faster than the CPU version.
Modal sound synthesis is a promising approach for real-time physically-based sound synthesis. A modal analysis is used to compute characteristic vibration modes from the geometry and material properties of scene objects. These modes allow an efficient sound synthesis at run-time, but the analysis is computationally expensive and thus typically computed in a pre-processing step. In interactive applications, however, objects may be created or modified at run-time. Unless the new shapes are known upfront, the modal data cannot be pre-computed and thus a modal analysis has to be performed at run-time. In this paper, we present a system to compute modal sound data at run-time for interactive applications. We evaluate the computational requirements of the modal analysis to determine the computation time for objects of different complexity. Based on these limits, we propose using different levels-of-detail for the modal analysis, using different geometric approximations that trade speed for accuracy, and evaluate the errors introduced by lower-resolution results. Additionally, we present an asynchronous architecture to distribute and prioritize modal analysis computations.
