AIR-SOAR: Intelligent Multi-Level Control
نویسندگان
چکیده
Autonomous systems must be able to deal with dynamic, unpredictable environments in real time. Our video describes a system for intelligent control of an airplane, within a realistic flight simulator (the Silicon Graphics flight simulator). The simulator allows asynchronous control of the plane’s throttle, ailerons, elevator and other control surfaces by an external system, and it provides limited asynchronous sensing of the plane’s motion. The result is a highly dynamic, real time domain in which models of the plane (and, potentially, other aircraft) are updated 20 times a second. Control of flight is complex. Unexpected events such as wind or turbulence must be responded to in a timely fashion. Further, identical control movements have different effects depending on the plane’s position and environmental conditions, making precise prediction of action effects difficult. The agent must also deal with delays in feedback from its actions, waiting for the plane to respond to changes in the control surfaces. The domain requires simultaneous execution of a range of tasks at different levels of complexity and granularity, from high level maneuvers like takeoff, landing and banked turns to low level tasks such as maintaining altitude, keeping the wings level and controlling the stick.
منابع مشابه
Intelligent Agents for Interactive Simulation Environments
Interactive simulation environments constitute one of today’s promising emerging technologies, with applications in areas such as education, manufacturing, entertainment and training. These environments are also rich domains for building and investigating intelligent automated agents, with requirements for the integration of a variety of agent capabilities, but without the costs and demands of ...
متن کاملA symbolic solution to intelligent real-time control
Autonomous systems must operate in dynamic, unpredictable environments in real time. The task of ying a plane is an example of an environment in which the agent must respond quickly to unexpected events while pursuing goals at di erent levels of complexity and granularity. We present a system, Air-Soar, that achieves intelligent control through fully symbolic reasoning in a hierarchy of simulta...
متن کاملControl of a Six-Legged Mobile Robot Using the Soar Cognitive Architecture
This paper describes the development of the Cognitive Robotic System (CRS) for studying intelligent and autonomous unmanned vehicles. This system uses a multi-threaded software system written using Java to integrate the Soar cognitive architecture with a sixlegged mobile robot and its sensors and motors. The results of experiments in which the Cognitive Robotic System was used to control the he...
متن کاملPogamut toolkit (demonstration)
All experiments using intelligent virtual agents, sooner or later, ask for a specific virtual environment that would fit their setup. Seeking such environment is a daunting task accompanied with the need for an appropriate agent adapter that provides infrastructure for mediation of virtual body senses and actions thereby enabling remote high-level agent control. This demo presents Pogamut toolk...
متن کاملA Soar-based Planning Agent for Gas-Turbine Engine Control and Health Management
Intelligent Agent technologies constitute an important stream of research in the Artificial Intelligence community. Their characteristics make them suitable for a variety of applications. In this paper, we investigate the use of Intelligent Agent technology in the field of Gas-Turbine Engine Control and Health Management. We present and test a Planning agent, developed to choose and apply appro...
متن کامل