Tuning Models of Pilot Tracking Behavior for a Specific Simulator Motion Cueing Setting

Abstract

This paper describes the preliminary results of an effort to compile data from a large number of studies that investigated the effects of variations in motion filter settings on pilot behavior. The main objective of this study is to formulate a set of mathematical rules that will allow for the tuning of behavioral pilot models to a certain motion cueing setting. To achieve this, data for different dependent measures such as tracking performance, pilot-vehicle system crossover frequencies, and pilot model parameters, taken from ten different experiments that considered pilot tracking behavior under varying rotational or translational motion cueing settings, has been combined. By checking the correlation of the variation in any of these dependent measures and parameters that quantify the applied variation in motion cueing, a number of consistent relations has been identified. The most consistent and clear effects that are found from this analysis are variations in some important dependent measures with the motion filter gain at 1 rad/s. Over the full range of motion filter gains at 1 rad/s from 0 to 1, a reduction in pilot visual gain of around 20% is observed with reducing motion filter gain, in combination with a 30% increase in the amount of visual lead equalization adopted by pilots.

Interactions of Outside Visual Cues and Motion Cueing Settings in Yaw Tracking

“…(6)), human operators decrease their visual lead time-constant τ L , while increasing their K pe , resulting in improved tracking performance. 11,16,17,30 Numerous studies have reported that these beneficial effects of motion feedback are suppressed when motion feedback is attenuated by high-pass filters. 3,16,17,31 In Ref.…”

“…11,16,17,30 Numerous studies have reported that these beneficial effects of motion feedback are suppressed when motion feedback is attenuated by high-pass filters. 3,16,17,31 In Ref. 17 an effort was made to quantify these effects, by deriving first-order prediction equations for the expected changes in human operator control parameters, as well as crossover frequencies and phase margins, from a database of a large number of previous studies.…”

“…3,16,17,31 In Ref. 17 an effort was made to quantify these effects, by deriving first-order prediction equations for the expected changes in human operator control parameters, as well as crossover frequencies and phase margins, from a database of a large number of previous studies. As the "predictor" variable the motion filter gain at 1 rad/s, symbolized as K S , was used.…”

“…3,31 The prediction equations of Ref. 17, here included as Eq. (7) to (12), will be used for comparison with our collected experiment data.…”

“…[11][12][13][14] Furthermore, it was found that as the quality of motion feedback degrades, by changes in the motion filter settings, human operators' tracking behavior approaches behavior as measured without motion feedback. 2,[15][16][17] The performance increase observed for the addition of motion feedback reduces, as the mismatch of the motion cues increases, which is attributed to a reduction in control gain and an increase in the amount of visual lead generated.…”