Showing papers on "Physics engine published in 2015"

Galileo: perceiving physical object properties by integrating a physics engine with deep learning

Abstract: Humans demonstrate remarkable abilities to predict physical events in dynamic scenes, and to infer the physical properties of objects from static images. We propose a generative model for solving these problems of physical scene understanding from real-world videos and images. At the core of our generative model is a 3D physics engine, operating on an object-based representation of physical properties, including mass, position, 3D shape, and friction. We can infer these latent properties using relatively brief runs of MCMC, which drive simulations in the physics engine to fit key features of visual observations. We further explore directly mapping visual inputs to physical properties, inverting a part of the generative process using deep learning. We name our model Galileo, and evaluate it on a video dataset with simple yet physically rich scenarios. Results show that Galileo is able to infer the physical properties of objects and predict the outcome of a variety of physical events, with an accuracy comparable to human subjects. Our study points towards an account of human vision with generative physical knowledge at its core, and various recognition models as helpers leading to efficient inference.

Chrono: An Open Source Multi-physics Dynamics Engine

Abstract: We provide an overview of a multi-physics dynamics engine called Chrono. Its forte is the handling of complex and large dynamic systems containing millions of rigid bodies that interact through frictional contact. Chrono has been recently augmented to support the modeling of fluid-solid interaction (FSI) problems and linear and nonlinear finite element analysis (FEA). We discuss Chrono’s software layout/design and outline some of the modeling and numerical solution techniques at the cornerstone of this dynamics engine. We briefly report on some validation studies that gauge the predictive attribute of the software solution. Chrono is released as open source under a permissive BSD3 license and available for download on GitHub.

Interactive real-time physics: An intuitive approach to form-finding and structural analysis for design and education

Abstract: Real-time physics simulation has been extensively used in computer games, but its potential has yet to be fully realized in design and education. We present an interactive 3D physics engine with a wide variety of applications. In common with traditional FEM, the use of a local element stiffness matrix is retained. However, unlike typical non-linear FEM routines elements forces, moments and inertia are appropriately lumped at nodes following the dynamic relaxation method. A semi-implicit time integration scheme updates linear and angular momentum, and subsequently the local coordinate frames of the nodes. A co-rotational approach is used to compute the resultant field of displacements in global coordinates including the effect of large deformations. The results obtained compare well against established commercial software. We demonstrate that the method presented allows the making of interactive structural models that can be used in teaching to develop an intuitive understanding of structural behaviour. We also show that the same interactive physics framework allows real-time optimization that can be used for geometric and structural design applications.

VR-OOS: The DLR's virtual reality simulator for telerobotic on-orbit servicing with haptic feedback

Abstract: The growth of space debris is becoming a severe issue that urgently requires mitigation measures based on maintenance, repair, and de-orbiting technologies. Such on-orbit servicing (OOS) missions, however, are delicate and expensive. Virtual Reality (VR) enables the simulation and training in a flexible and safe environment, and hence has the potential to drastically reduce costs and time, while increasing the success rate of future OOS missions. This paper presents a highly immersive VR system with which satellite maintenance procedures can be simulated interactively using visual and haptic feedback. The system can be used for verification and training purposes for human and robot systems interacting in space. Our framework combines unique realistic virtual reality simulation engines with advanced immersive interaction devices. The DLR bimanual haptic device HUG is used as the main user interface. The HUG is equipped with two light-weight robot arms and is able to provide realistic haptic feedback on both human arms. Additional devices provide vibrotactile and electrotactile feedback at the elbow and the fingertips. A particularity of the realtime simulation is the fusion of the Bullet physics engine with our haptic rendering algorithm, which is an enhanced version of the Voxmap-Pointshell Algorithm. Our haptic rendering engine supports multiple objects in the scene and is able to compute collisions for each of them within 1 msec, enabling realistic virtual manipulation tasks even for stiff collision configurations. The visualization engine ViSTA is used during the simulation to achieve photo-realistic effects, increasing the immersion. In order to provide a realistic experience at interactive frame rates, we developed a distributed system architecture, where the load of computing the physics simulation, haptic feedback and visualization of a complex scene is transferred to dedicated machines. The implementations are presented in detail and the performance of the overall system is validated. Additionally, a preliminary user study in which the virtual system is compared to a physical test bed shows the suitability of the VR-OOS framework.

Modelling granular soil behaviour using a physics engine

Abstract: A physics engine is a software library used in the film and computer games industries to realistically animate a physical system. In this paper it is shown that particulate media can be faithfully modelled using a rigid body physics engine, thereby providing a viable alternative to the discrete element method codes currently used in the field of geomechanics. An overview of the simulation method implemented in the widely used Box2D physics engine is provided, and it is shown that this tool can successfully capture the critical state response of granular media, using particles modelled as randomly shaped polygons.

Simulation of granular soil behaviour using the bullet physics library

Abstract: A physics engine is computer software which provides a simulation of certain physical systems, such as rigid body dynamics, soft body dynamics and fluid dynamics. Physics engines were firstly developed for using in animation and gaming industry ; nevertheless, due to fast calculation speed they are attracting more and more attetion from researchers of the engineering fields. Since physics engines are capable of performing fast calculations on multibody rigid dynamic systems, soil particles can be modeled as distinct rigid bodies. However, up to date, it is not clear to what extent they perform accurately in modeling soil behaviour from a geotechnical viewpoint. To investigate this, examples of pluviation and vibration-induced desification were simulated using the physics engine called Bullet physics library. In order to create soil samples, first, randomly shaped polyhedrons, representing gravels, were generated using the Voronoi tessellation approach. Then, particles were pluviated through a funnel into a cylinder. Once the soil particles settled in a static state, the cylinder was subjected to horizontal sinusoidal vibration for a period of 20 seconds. The same procedure for sample perparation was performed in the laboratory. The results of pluviation and vibration tests weere recorded and compared to those of simulations. A good agreement has been found between the results of simulations and laboratory tests. The findings in this study reinforce the idea that physics engines can be employed as a geotechnical engineering simulation tool.

Controlling physical toys using a physics engine

Abstract: Methods and systems for controlling physical toys using a physics engine are described. In an embodiment, a physics engine within an interactive software experience is used to model the motion of a virtual object in a virtual environment. The output of this modelling is then imposed on a physical toy which corresponds to the virtual object such that the motion of the physical toy in the real world more closely matches the motion of the virtual object in the virtual environment. In various examples, the modelling is imposed through control signals which are generated based on output of the physics engine and used to control actuators within the physical toy to change the motion of at least a part of the toy.

GPU Accelerated Real-Time Collision Handling in Virtual Disassembly

Abstract: Previous collision detection methods for virtual disassembly mainly detect collisions at discrete time intervals, and use oriented bounding boxes to speed up the process. However, these discrete methods cannot guarantee no penetration occurs when the components move. Meanwhile, because some of the components are embedded into each other, these components cannot be separated in the subsequent process. To solve these problems, we propose an approach for real-time collision handling by utilizing the computational power of modern GPUs. First we present a novel GPU-based collision handling framework for virtual disassembly. Second we use a collision-streams based continuous collision detection to guarantee no collision missed. Finally we introduce a triangle intersection detection algorithm to solve the problem that collision cannot be detected when the components are embedded into each other at the initial configuration. The experimental results show that our method can improve the overall performance of collision detection and achieve real-time simulation.

UnityMol: interactive and ludic visual manipulation of coarse-grained RNA and other biomolecules

Abstract: We present a general software architecture to carry out interactive molecular simulations in a game engine environment. Our implementation is based on the UnityMol framework and the HireRNA physics engine. With UnityMol, we pursue the goal to create an interactive virtual laboratory enabling researchers in biology to visualize biomolecular systems, run simulations and interact with physical models and data. Similarly, UnityMol enables game designers to build scientifically accurate molecular scenarios. We discuss four case studies, from simulation setup via immersive experiments, force-induced unfolding of RNA to teaching and collaborative research applications. Visual effects enrich the dynamic and immersive aspects. We combine an appealing visual feedback with a set of analysis features to extract information about properties of the fascinating biomolecular systems under study. Access to various input devices enables a natural interaction with the simulation.

Maintain and Improve Mental Health by Smart Virtual Reality Serious Games

Abstract: Serious games for mental health is seen as the groundwork for assistive technology to maintain and improve mental health. We present a technical system layout we partly implemented for demonstration purposes and highlight vision-based perception and manipulation capabilities. These include physical interactions employing artificial general intelligence in virtual reality applications. We employ hand gesture tracking, as well as an Oculus Rift integrated gaze and eye tracking system. The resulting serious games should eventually cover daily life activities, which we additionally monitor. The dynamic and contextual modelling of obstacles are central issues, and capabilities required for serious games include knowledge about the 3D world. Such knowledge include gaze and hand sensors interpretations for multimedia information extraction in causal relationships. Towards this goal, we envision to make use of virtual reality with a physics engine (rigid and soft body dynamics including collision detection) for the observed objects. We also exploit semantic networks to enable the machine to filter information and infer ongoing complex events including hidden BDI (beliefs, desires, intentions) variables. We see this combination of employed technology as the relevant groundwork for reaching human-level general intelligence and to enable real-world applications. Future applications and user groups we target on include dementia patients.

A Case-Based Approach to Mobile Push-Manipulation

Abstract: The complexity of the potential physical interactions between the robot, each of the pushable objects, and the environment is vast in realistic mobile push-manipulation scenarios. This makes it non-trivial to write generic analytical motion and interaction models or tune the parameters of physics engines for every robot, object, and environment combination. We present a case-based approach to push-manipulation that allows the robot to acquire, through interaction and observation, a set of discrete, experimental, probabilistic motion models (i.e. probabilistic cases) for pushable passively-mobile real world objects. These probabilistic cases are then used as building blocks for constructing achievable push plans to navigate the object of interest to the desired goal pose as well as to potentially push the movable obstacles out of the way in cluttered task environments. Additionally, incremental acquisition and updating of the probabilistic cases allows the robot to adapt to the changes in the environment, such as increased mass due to loading of the object of interest for transportation purposes. The purely interaction and observation driven nature of our method makes it robot, object, and environment (real or simulated) independent, as we demonstrate through validation tests in a real world setup in addition to extensive experimentation in simulation.

Growing and Evolving Vibrationally Actuated Soft Robots

Abstract: Designing soft robots is difficult, time-consuming, and non-intuitive. Soft robot design faces two main challenges: structure and control. This research uses generative encodings to grow structures and a vibrational mechanism to control locomotion. In this paper, we demonstrate the ability to successfully evolve soft robots that can move when vibrated. Soft bodies are grown through a grammatical process and simulated in the Bullet physics engine. We also briefly outline a method of evolving scalable solutions that we are currently investigating. It should be capable of generating soft robots of various sizes that can move when vibrated.

Comparison of Multibody Dynamics Solver Performance: Synthetic Versus Realistic Data

Abstract: In the area of robotics simulation, multibody dynamics plays an important role in designing and controlling robots, especially when the robot contacts the environment. Contacts give rise to non-penetration and friction constraints, which are nonsmooth and nonlinear. One way to simulate such systems is through the use of a discrete-time multibody dynamics model in the form of a nonlinear complementarity problem (NCP), for which, finding a solution is known to be NP-hard [1]. In situations where analytical solutions don’t exist, a suite of numerical solutions accessible through a benchmarking framework is useful to fairly evaluate performance of different computer algorithms. However, many algorithm designers don’t have easy access to test data from physical simulators. Under such circumstances, randomized data are used to test the performance of solution algorithms.In this paper, we present our Benchmark Problems of Multibody Dynamics (BPMD) framework and database, with the data sets from different physics engines, as a benchmarking platform. Then we compare the performance of several solvers on synthetic data and simulation data, to show the superiority of testing solution algorithms with simulation data over testing with synthetic data. We will show that algorithm tested only on synthetic data often fail to solve problem obtained from physics simulations, to demonstrate the benefit of BPMD database.Copyright © 2015 by ASME

GPU ray-traced collision detection for cloth simulation

Abstract: We propose a method to perform collision detection with cloths with ray-tracing. Our method is able to perform collision detection between cloths and volumetric objects (rigid or deformable) as well as collision detection between cloths (including auto-collision). Our method casts rays between objects to perform collision detection, and an inversion-handling algorithm is introduced to correct errors introduced by discrete simulations. GPU computing is used to improve the performances. Our implementation handles scenes containing deformable objects at an interactive frame-rate, with collision detection lasting a few milliseconds.

Evolving Honeycomb Pneumatic Finger in Bullet Physics Engine

Abstract: Soft robot is becoming a current focus for its inherently compliance and human-friendly interacting with the real world. But most soft robots are designed and fabricated with intuition and empiricism only, which lack of systematic assessment such as force and deformable analysis before fabricating. Before choosing proper soft materials and processing the craft, the experiments of the soft robots can hardly be set up. In this paper, we construct a model of Honeycomb Pneumatic Finger (HPF) with honeycomb pneumatic network embedded which overcomes the shortcoming of embedded rectangular unit. In the meantime, a pressure analysis model is built for the purpose of physical simulation. Based on the model, without choosing any real materials and fabricating, we focus on exploring the correlations between the pressure and the geometrical shapes in physics simulation. Furthermore, we construct a virtual hand consisting of one rigid palm and four HPFs which is simulated with Bullet Physics Engine. By changing the pressure of the corresponding honeycomb units of each finger, the hand can grasp a ball smoothly and lift it up. At last, the deviation between the mathematical analysis and the physical simulation is discussed.

Using game physics engines for hardware-in-the-loop material flow simulations: benefits, requirements and experiences

Abstract: Object of this paper is the proposal of using existing physics engines mainly deployed in video games and movie animations for simulating material flow applications with a high degree of reality. Since the actual control systems are supposed to be tested and verified using this simulation it has to possess real-time capabilities for usage with a hardware-in-the-loop simulation setup. This paper illustrates why such a simulation is necessary concerning the considered application. In addition, the required work items to achieve this approach are presented. During the realization, encountered problems and obstacles are commented on as well.

MecaCell: an Open-source Efficient Cellular Physics Engine

Abstract: We present an open source physics engine specialised for multi-cellular artificial organisms simulations. It is computationally efficient in comparison to gas-based and finite element models and more realistic than standard mass-spring-damper systems.

Application of Image Processing and Virtual Reality Technologies in Simulation of Laparoscopic Procedures

Abstract: In this paper we focus on developing the simulators of laparoscopic surgery procedures. We propose expanding the role of video channel, which allows to reduce the mechatronic systems providing information about the spatial position of instruments. The mechatronic structure is indispensable in order to achieve the effect of haptic feedback, however in the proposed solution it is much simpler, than as in the case of currently occurring solutions. In proposed issue, mechatronic consists of actuators with a simple control software only. Entire perception (data acquisition) for the realization of haptic feedback is implemented using the video channel. Moreover, haptic feedback effect is supported by the dynamics calculations made in the graphics software. While calculating of the movement of the objects, their collisions and associated with that deformations, the physics engine purveyed current value of forces, torques, speed or acceleration to haptic feedback fittings.

Dynamic 3D interaction using an optical See-through HMD

Abstract: We propose a system that enables dynamic 3D interaction with real and virtual objects using an optical see-through head-mounted display and an RGB-D camera. The virtual objects move according to physical laws. The system uses a physics engine for calculation of the motion of virtual objects and collision detection. In addition, the system performs collision detection between virtual objects and real objects in the three-dimensional scene obtained from the camera which is dynamically updated. A user wears the device and interacts with virtual objects in a seated position. The system gives users a great sense of reality through an interaction with virtual objects.

Efficient Collision Detection for Industrial Robot Simulation System

Abstract: As the result of an increasing number of industrial robots, the simulation of industrial robot has become a very active research field. Collision detection for industrial robot simulation system is an essential part. In this paper, we present two kinds of collision detection algorithms, namely, internal collision detection algorithm and external collision detection algorithm. These two kinds of algorithms are based mainly on the bounding volume technology. Internal collision detection algorithm isi??applicable toi??the collisions between robotic links and external collision detection algorithm isi??applicable toi??the collisions between robotic links and its surrounding obstacles. In addition, we also present two examples of simulation results. The examples prove that the collision detection module can satisfy the collision detection requirement of the industrial robot simulation system.

A method for 3D rock fracturing simulation based Havok

Abstract: In order to solve the incompatibility between real-time property and realism, we introduce a method for 3D rock fracturing simulation based on the Havok Physical Engine in this paper. Firstly, during the model preprocessing stage, the seed points are inserted inside the model and the 3D Voronoi fragments are generated as many as the seed points. At the same time, the breakable linear constraints between the fragments are established. Then, during the dynamics simulation stage, the rigid body fracturing simulation is achieved by processing the dynamics collision such as the collision detection and collision response. In this paper, the method is applied in the rock model simulation with a good effect. The dynamics property of rigid body is ideal. The simulation presents both perfect physical reality and real-time property. As shown in the experimental results, the fracturing algorithm has good universality, which can be applied to the simulation of rock, wall, plaster and other materials. The simulation frames can reach over 60 fps.

New geometric algorithms and data structures for collision detection of dynamically deforming objects

Abstract: Any virtual environment that supports interactions between virtual objects and/or a user and objects, needs a collision detection system to handle all interactions in a physically correct or plausible way. A collision detection system is needed to determine if objects are in contact or interpenetrates. These interpenetrations are resolved by a collision handling system. Because of the fact, that in nearly all simulations objects can interact with each other, collision detection is a fundamental technology, that is needed in all these simulations, like physically based simulation, robotic path and motion planning, virtual prototyping, and many more. Most virtual environments aim to represent the real-world as realistic as possible and therefore, virtual environments getting more and more complex. Furthermore, all models in a virtual environment should interact like real objects do, if forces are applied to the objects. Nearly all real-world objects will deform or break down in its individual parts if forces are acted upon the objects. Thus deformable objects are becoming more and more common in virtual environments, which want to be as realistic as possible and thus, will present new challenges to the collision detection system. The necessary collision detection computations can be very complex and this has the effect, that the collision detection process is the performance bottleneck in most simulations. Most rigid body collision detection approaches use a BVH as acceleration data structure. This technique is perfectly suitable if the object does not change its shape. For a soft body an update step is necessary to ensure that the underlying acceleration data structure is still valid after performing a simulation step. This update step can be very time consuming, is often hard to implement and in most cases will produce a degenerated BVH after some simulation steps, if the objects generally deform. Therefore, the here presented collision detection approach works entirely without an acceleration data structure and supports rigid and soft bodies. Furthermore, we can compute inter-object and intraobject collisions of rigid and deformable objects consisting of many tens of thousands of triangles in a few milliseconds. To realize this, a subdivision of the scene into parts using a fuzzy clustering approach is applied. Based on that all further steps for each cluster can be performed in parallel and if desired, distributed to different GPUs. Tests have been performed to judge the performance of our approach against other state-of-the-art collision detection algorithms. Additionally, we integrated our approach into Bullet, a commonly used physics engine, to evaluate our algorithm. In order to make a fair comparison of different rigid body collision detection algorithms, we propose a new collision detection Benchmarking Suite. Our Benchmarking Suite can evaluate both the performance as well as the quality of the collision response. Therefore, the Benchmarking Suite is subdivided into a Performance Benchmark and a Quality Benchmark. This approach needs to be extended to support soft body collision detection algorithms in the future.

A Simple Filtering Algorithm for Continuous Collision Detection Using Taylor Models

Abstract: A huge number of potentially colliding triangles go to the succeeding narrow stage of continuous collision detection, even though a broad culling technique such as bounding volume hierarchies is applied. This heavily burdens the elementary collision tests in a collision detection algorithm and affects the performance of the entire pipeline, especially for fast moving or deforming objects. We present a low-cost filtering algorithm using Taylor Models. The experiments show that our algorithm can significantly reduce the number of elementary collision tests that occur in the narrow stage of collision detection.

Optimizing on real-time fluid 3D effect in mobile environment

Abstract: Simulating fluids on screen has been getting many spotlights because it is one of the most challenging tasks in computer science area. So we developed a real-time fluid simulation on android platform with SPH fluid algorithm and Bullet physics engine, and then rendered it with OpenGL|ES. Because of the limitation of smartphone environment, the simulation does not show a good performance. We suggest three methods to improve the performance of the demo. Firstly, we enable parallel computing with ARM NEON to accelerate mathematical computation. Secondly, we simplify the SPH algorithm with particle-forming technique. Finally, we optimize rendering skill with SSFR technique. We carry out experiments to validate the methods we suggest. The experiments results show that the first method improved the performance by about 26% to 40%, the second method improved the performance by about 50%, and the third method helped us represent a more smooth and natural fluid surface. Therefore, these methods are proved to be able to improve the performance of the real-time 3D fluid effect simulation significantly.

Sphere encapsulated oriented-discrete orientation polytopes (S-dop) collision culling for multi-, rigid body dynamic

Abstract: This paper discusses on sphere encapsulated oriented-discrete orientation polytopes (therefore will be referred to as S-Dop) collision culling for multiple rigid body simulation. In order to improve performance of the whole simulation system, there are available options in sacrificing the accuracy over speed by using certain approximation techniques. The aim of this research is to achieve excellent performance through implementation of suitable culling technique, without jeopardizing the resulting behavior so that the simulation will still be physically plausible. The basic idea is to identify the highly probable pairs to collide and test the pair with a more accurate collision test in broad-phase collision detection, before the pair is passed to a more costly stage. Results from the experiments showed that there are a number of ways to implement the sphere encapsulated or-Dops (S-Dop) collision culling on a multiple rigid body simulation depending on the level of performance needed.

The ACT-R Unity Interface: Integrating ACT-R with the Unity Game Engine

Abstract: Cognitive architectures are computational frameworks that support the development of computational models of human cognitive processes. They have typically been used to advance our understanding of human cognition in specific task environments; however, they have also been used to support the development of a variety of intelligent systems and agents (e.g., cognitive robots). There are a broad range of reasons to motivate the effort to integrate cognitive architectures with virtual environments. These include the development of intelligent virtual characters for the purposes of training simulations, enhanced gameplay experiences and the modelling of user/gamer behaviour. Virtual environments also support simulations of actual real-world environments (e.g., using physics engines and advanced lighting models) that can be used to perform computational simulations into embedded, extended, and embodied cognition. In the current report, we describe the effort to develop an integration framework that enables the ACT-R cognitive architecture to be used in conjunction with the Unity game engine. The resulting framework, referred to as the ACT-R Unity Integration (ACT-R UI) framework, enables individual ACT-R models to control the behaviour of virtual non-player characters that inhabit 3D virtual environments built on top of Unity. We first provide an overview of the ACT-R architecture and the Unity game engine. We describe the key features of both systems and discuss why they provide such a compelling target for integration. We then go on to describe the nature of the integration solution itself. We outline the extensions to the ACT-R architecture that enable ACT-R models to exchange information with Unity, and we also present the Unity components that enable virtual characters to be controlled or influenced by ACT-R models. Finally, we provide a concrete example of the use of the ACT-R UI. In particular, we show how an ACT-R model can be used to control the behaviour of a virtual robotic character that inhabits a Unity-based virtual environment.

Novel force feedback interface for improving human perception in laparoscopic surgical trainer

Abstract: This paper considers interaction of the laparoscopic instrument with virtual objects simulated in minimally invasive surgery trainer, which is a complex device which requires a synthesis between visual and haptic information. To compute the point-to-point interaction forces between the laparoscopic instrument and human internal organs, the software graphic data processing algorithms can be implemented. Inspired by the idea of Haptic Virtual Objects and virtual reality architecture, incorporated visual and haptic feedback, we employ virtual reality (VR) software for detection of collision between soft and rigid body, and soft body deformation. VR physics engine apply a force to each vertex of a triangle tessellated mesh, so as a result we obtain from VR physics rendering engine ready to use force vectors. To obtain that, first we create computer-generated Haptic Virtual Objects (HVO), which can be touched and manipulated with laparoscopic instruments. Next, we employ specific rendering engine working on physical interactions. In proposed software (written in C++) we employ HVO and OGRE 3D graphic rendering and Bullet Physics engines. It calculates a force feedback for yaw, pitch and insertion axes of motion. Force rendering algorithm considers point-based interactions, and the force rendering loop update rate is about 100Hz.

A multi-scale, physics engine-based simulation of cellular migration

Abstract: This research paper describes the design and prototyping of a simulation tool that provides a platform for studying how behavior of proteins in the cell membrane influences macro-level, emergent behaviors of cells. Whereas most current simulation tools model cells as homogeneous objects, this new tool is designed to modularly represent the cell's complex morphology and the varying distribution of proteins across the membrane. The simulation tool uses a physics engine to manage motion and collisions between objects. It also represents dynamic fluid environments, experimental surfaces, attachment bonds and interactions between the dynamically changing cell surface proteins. The prototype tool is described along with proposals for its use and further development.