Nico Hempe

Advanced Sensor Simulation In Virtual Testbeds: A Cost-Efficient Way to Develop and Verify Space Applications

Developing and testing components for space applications is a cost-intensive and timecritical process. Therefore, prospective missions and projects in space business will mostly be carried out in international cooperation, e.g. between the American space agency NASA and the German space agency DLR. A major challenge is the distributed development of components, implementation of algorithms and testing, as expensive hardware cannot be provided at each research facility. To resolve this situation, so called ’Virtual Testbeds’ (VT) are currently being developed. VTs provide a comprehensive and integrated simulation of the components as well as of the entire target systems and thus allow to perform a major part of the required testing without the physical hardware. In this paper we present the ’Virtual Testbed’ concept, which is able to virtualize complete mission scenarios. A focus is laid on the highly realistic sensor simulation, as they are a very important part in many virtual testbed applications supporting the envisaged iterative development and validation process. The implemented sensor simulation framework is accurate, versatile and exible enough to help to eliminate systematical errors not only during the virtual prototyping phase. Developers also use it for control code debugging, subsystem- and system validation in the implementation and test phases. This results in shorter development times, while simultaneously decreasing the developing costs dramatically.

Geometric Interpretation and Optimization of Large Semantic Data Sets in Real-Time VR Applications

Current real-time VR applications are based on well-defined digital representations of the environment. In order to render a realistic looking environment with good performance, artists and developers with specific expertise are indispensable to create optimized data. However modern applications, especially those incorporating data from geo information (GIS) or product data management (PDM) systems, need to be able to use unrefined data without offline conversion or loss of render performance. In this paper we present an extensible object oriented graph database, which further embraces the paradigm of object orientation by incorporating the simulation functionality into the database itself. Whole scene descriptions including all functionalities can be described by one single database. Optimization techniques will be introduced, which are automatically applied to the simulation data, in order to extract a render-friendly structure. Specific semantic objects can be interpreted by the render framework to enhance the simulation, in both function and visual representation.Copyright

A Semantics-Based, Active Render Framework to Realize Complex eRobotics Applications with Realistic Virtual Testing Environments

Multi-domain VR simulation systems provide a framework to bring together various modules in order to achieve the desired goals. Due to different data structure needs regarding rendering and simulation tasks, common systems typically focus on one of these areas. In complex technical or mechatronical systems like mobile robotic applications, rendering capabilities are mostly neglected. Hence, the used models are designed with minimal effort resulting in purely functional virtual environments with low visual quality. The novel field of eRobotics is currently being developed as an intuitively applicable eService related to VR in advanced robotic and engineering applications. In this contribution, we will introduce novel system structures and techniques, which meet the requirements to realize a holistic eRobotics development support tool, combining complex simulations with state-of-the art rendering techniques. Semantic world models build the ideal basis in order to combine both areas. Readily available data like provided by geo information services (GIS) can be used as input for the creation of close-to-reality virtual testing environments with minimal effort. Additionally, the close interplay of simulation and rendering modules allows making the transition from a passive render component towards an active one, further improving the accuracy and performance of non-rendering-related tasks like optical sensor simulations.

Combining realistic virtual environments with real-time sensor simulations for close-to-reality testing and development in eRobotics applications

Today, Virtual Reality (VR) simulation technology is a well-known field of engineering and widely applied in research and in the industry. In particular, the simulation of various sensors is an essential task in robotics-related contexts and part of many current multi-domain VR simulation systems. In this contribution, we present the novel concept of eRobotics, which combines rendering and simulation tasks to not only provide attractive virtual environments, but also helps to develop and test virtual prototypes with accurate, real-time optical sensor simulations like digital camera systems, laser range scanners (LiDAR) and time-of-flight (ToF) cameras, capable to meet real world conditions. Practical applications will demonstrate the advantages of these concepts in multiple robotics-related domains, which help to make the developed approaches and systems more robust before testing them in real world setups.

A flexible model for real-time crowd simulation

This paper will introduce the generic concept of a multi-agent based crowd simulation prototype. The prototype consists of many distinct components that contribute to the system as a whole. Based on the criteria for a human agent model, the agents module-based, layered architecture is introduced. Subsequently, a closer examination of each module reveals a detailed insight into the agents architecture. The examined modules are: The agents finite state machine, its steering behavior, its locomotion model, the path planner and the messaging capability. Finally, with respect to the simulations general architecture, the optimization techniques to further improve the simulations runtime performance are discussed. This is especially important, since the simulation should run at interactive frame rates and allow the simulation of as many agents as possible for scalability reasons. The investigated optimization techniques are multi-threading and cell space partitioning. The simulation is implemented in our 3D simulation system VEROSIM, which also handles 3D graphics to visualize the results in real-time.

Connecting the areas of eLearning, virtual training and engineering in multidisciplinary eRobotics applications

The recent trend of Virtual Reality (VR) already found its way into entertainment and scientific applications. While current entertainment domains fully exploit the technical possibilities of modern graphics hardware and state of the art rendering approaches, many VR simulation system in scientific contexts put the focus on numerical results with rather functional representations. The objective of the novel research field of eRobotics is to effectively use electronic media to achieve the best possible advancements in the development of robotics-related applications. Corresponding eRobotics systems take the step from rather functional scientific simulations to attractive demonstrators using VR technologies, which can not only be applied in motivating eLearning scenarios but also in full-featured engineering applications in the context of virtual testbeds and cyber-physical systems. In this paper, we will introduce concepts and system structures that provide new opportunities for the development of a complex but comprehensive multidisciplinary development support tool. A broad range of practical applications in different domains will demonstrate the benefits of connecting the areas of eLearning, virtual training and engineering in a holistic eRobotics system.

Real-Time Rendering Approaches for Dynamic Outdoor Environments

eRobotics systems should provide new opportunities through the combination of complex simulations at scientific level with attractive real-time computer graphics. An important precondition for the realization of such a unifying approach is to provide possibilities that enable the creation and rendering of realistic virtual environments with minimal effort. These attractive virtual environments not only increase the user acceptance, they are also well suited to be applied as virtual test grounds, which enable close-to-reality testing of novel approaches and virtual prototypes under defined but dynamic conditions.

Summary, Conclusion and Outlook

Over the last decades, VR simulation systems have become an omnipresent part in many application areas ranging from edutainment over virtual training up to engineering and virtual prototyping. Each task and each domain demand specific features, data structures and functionalities, which are difficult to combine in an efficient way. This is particularly true in the areas of simulation and rendering, which lead to the new idea of a unifying approach based on the principles of the novel research field of eRobotics.

Concept and Realization of a Multi-Domain VR Simulation System

It is still a challenge to develop a comprehensive VR Simulation System that has the capabilities to serve as a development basis in a wide range of domains. When investigating multiple application areas, a large number of specific software components and libraries are required to fulfill all desired goals. By the definition of flexible but comprehensive and standardized interfaces, system structures and communication possibilities, the number of applied frameworks, which can work together in a system without affecting each other in a negative way, can drastically be increased.

Combining Complex Simulations with Realistic Virtual Testing Environments The eRobotics-Approach for Semantics-Based Multi-domain VR Simulation Systems

Today Virtual Reality (VR) simulation technology is a well-known field of virtual training and engineering and widely applied in research and in the industry. Multi-domain VR simulation systems cover multiple technical and visual aspects not limited to a single task or domain. While current systems mostly neglect the rendering component and provide purely functional graphics and simple virtual environments, we present the concepts of eRobotics and matching system structures to combine complex simulations and realistic virtual environments in a holistic VR simulation system. These environments not only provide attractive visual presentations, they also help to realize close-to-reality testing of virtual prototypes and positively affect the accuracy and performance of simulated components like optical sensors.