Virtual World Visualization for an Autonomous Underwater Vehicle

A critical bottleneck exists in Autonomous Underwater Vehicle (AUV) design and development. It is tremendously difficult to observe, communicate with and test underwater robots, because they operate in a remote and hazardous environment where physical dynamics and sensing modalities are counterintuitive. An underwater virtual world can comprehensively model all necessary functional characteristics of the real world in real time. This virtual world is designed from the perspective of the robot, enabling realistic AUV evaluation and testing in the laboratory. 3D real-time graphics are our window into the virtual world, enabling multiple observers to visualize complex interactions. A networked architecture enables multiple world components to operate collectively in real time, and also permits world-wide observation and collaboration with other scientists interested in the robot and virtual world. INTRODUCTION This paper describes the software architecture of an underwater virtual world for an autonomous underwater robot, emphasizing the importance of 3D real-time visualization in all aspects of the design process. Multiple component models provide interactive real-time response for robot and human users. Theoretical development stresses a scalable distributed network approach, interoperability between models, physics-based reproduction of real-world response, and compatibility with open systems approaches. Logical network connectivity of physical interactions is provided using the IEEE standard Distributed Interactive Simulation (DIS) protocol. Implementation of the underwater virtual world and autonomous robot are tested using the NPS AUV Phoenix (Figure 1). Documentation and source code are available electronically (Brutzman 94). Current work includes adapting hydrodynamics and controls coefficients for other submersibles. Future work includes porting graphics and network connectivity to the nascent Virtual Reality Modeling Language (VRML), permitting Internet-based rendering on any computer compatible with the World-Wide Web (WWW). To be presented at the IEEE Oceanic Engineering Society Conference OCEANS 95, San Diego California, October 10-12 1995. Figure 1. NPS AUV shown in test tank (Torsiello 94). MOTIVATION Underwater robots are normally called Autonomous Underwater Vehicles (AUVs), not because they are intended to carry people but rather because they are designed to intelligently and independently convey sensors and payloads. AUVs must accomplish complex tasks and diverse missions while maintaining stable physical control with six spatial degrees of freedom. Little or no communication with distant human supervisors is possible. When compared to indoor, ground, airborne or space environments, the underwater domain typically imposes the most restrictive physical control and sensor limitations upon a robot. Underwater robot design requirements therefore motivate this examination. Considerations and conclusions remain pertinent as worst-case examples relative to other environments. A large gap exists between the projections of theory and the actual practice of underwater robot design. Despite a large number of remotely operated submersibles and a rich field of autonomous robot research results, few AUVs exist and their capabilities are limited. Cost, inaccessibility and scope of AUV design restrict the number and reach of players involved. Interactions and interdependencies between hardware and software component problems are poorly understood. Testing is difficult, tedious, infrequent and potentially hazardous. Meaningful evaluation of results is hampered by overall problem complexity, sensor inadequacies and human inability to directly observe the robot in situ. Potential loss of