AIAA-94-1267-CP Virtual Environments for Telerobotic Shared Control

Traditional teleoperation depends solely on operator skill for effective task completion. This dependency limits the class of operations which were suitable for teleoperation. In many work situations today, external and operational requirements necessitate a more efficient operational scenario. The potential environments are more complex and dangerous and require integrated safety systems. One way to enable more effective control is to share control between the human operator and an intelligent computing system. This paper discusses the requirements and an implementation of a virtual environment for telerobotic shared control. Efficient, effective and safe system operation depends on the operator's ability to make wise decisions. In order to make wise decisions the operator needs a clear understanding of the operational requirements, and an understanding of the external factors which affect the operation requirements. A virtual environment built upon a shared control interface can provide much of the information required for safe operation. The virtual world allows the operator to clearly visualize the task and provides feedback that is not otherwise available. The shared control interface verifies operational demands against system constraints. The virtual environment provides for a practice and training environment where mistakes can be made. Only after an operation has been cleared for execution does physical motion take place. Proper linking of human input with computer controlled heuristics greatly increases the safety level of the system. The knowledge base that the computing system uses to perform decision making is called the World Model. Visual information from the World Model is displayed graphically as rendered, shaded, texture mapped animated polygons. The core of the shared control system is the computational engine that maintains, displays, visualizes and controls interaction with the World Model. The computational engine can be interrogated at several different levels. During operation: multilevel interaction defined by the program (mouse, keyboard, multi-modal input devices). The visualization system offers a high interactive operating environment, enabling users to change objects, display attributes, and elements of the WorldModel in real time. High speed and optimal interaction with external input and teaching devices are required for effective interaction. Traditional teleoperation allows the user to effect immediate changes in manipulator configurations via an input device(s). Anything from a simple joystick to a force reflecting master have been utilized. The virtual environment uses these input tools and more advanced tools such as datagloves to direct user input into the graphics environment. This enables the use of the graphics environment for training, simulation, design interactions and will allow the WorldModel to interact and interpret operator commands.