Multi-Loop Identification of Pilot Central Visual and Vestibular Motion Perception Processes

Abstract

When controlling an aircraft, a pilot generates control inputs based on a variety of perceived cues. Flight displays or the changing outside scenery provide visual cues, the motion of the aircraft provides vestibular cues. Detailed knowledge about the use of motion cues in control tasks is of great interest in the fields of aircraft, spacecraft and simulator (cockpit-) design, simulator fidelity research, but also in research on human physiology. This paper describes the results of a study on the pilot’s use of central visual and vestibular motion cues in manual control tasks. Simulator experiments were performed where subjects had to control a double-integrator system about the roll-axis of the simulator. Motion cues were provided by a compensatory central display and physical motion of the simulator. A reference signal as well as a disturbance signal was presented to the pilot simultaneously, allowing for multi-loop identification of separate response functions for the central visual and vestibular system. A multi-channel pilot model was fitted to the estimated response functions, allowing for the determination of distinct time delays for both modalities of perception. Reduction of magnitude of the input signals created two tasks with the characteristics of a target following and a disturbance compensating task. Both tasks were presented with full motion inputs and reduced motion inputs to the actuators of the simulator, to study in more detail how responses are generated by the vestibular system. A smaller time delay for the response of the vestibular system than for the central visual system was found. As a result of the reduction of motion, the time delay and crossover frequency of the vestibular response was found to increase, while performance was found to decrease. Higher crossover frequencies for the central visual response and more control activity was found for disturbance compensating tasks than for target following tasks.