Modeling the Human Visuo-Motor System for Remote-Control Operation

Abstract

Successful operation of a teleoperated miniature rotorcraft relies on capabilities including guidance, trajectory following, feedback control, and environmental perception. For many operating scenarios fragile automation systems are unable to provide adequate performance. In contrast, human-in-the-loop systems demonstrate an ability to adapt to changing and complex environments, stability in control response, high level goal selection and planning, and the ability to perceive and process large amounts of information. Modeling the perceptual processes of the human operator provides the foundation necessary for a systems based approach to the design of control and display systems used by remotely operated vehicles. In this work we consider flight tasks for remotely controlled miniature rotorcraft operating in indoor environments. Operation of agile robotic systems in three dimensional spaces requires a detailed understanding of the perceptual aspects of the problem as well as knowledge of the task and models of the operator response. When modeling the human-in-the-loop the dynamics of the vehicle, environment, and human perception-action are tightly coupled in space and time. The dynamic response of the overall system emerges from the interplay of perception and action. The main questions to be answered in this work are: i) what approach does the human operator implement when generating a control and guidance response? ii) how is information about the vehicle and environment extracted by the human? iii) can the gaze patterns of the pilot be decoded to provide information for estimation and control? In relation to existing research this work differs by focusing on fast acting dynamic systems in multiple dimensions and investigating how the gaze can be exploited to provide action-relevant information. To study human-in-the-loop systems the development and integration of the experimental infrastructure is described. Utilizing the infrastructure, a theoretical framework for computational modeling of the human pilot’s perception-action is proposed and verified experimentally. The benefits of the human visuo-motor model are demonstrated through application examples where the perceptual and control functions of a teleoperation system are augmented to reduce workload and provide a more natural human-machine interface.