Ergonomics Experiment Research on Visual Characteristics of Head-Up-Display Failure Warning
Authors: Sun guoqiang, Xu wu
Abstract: With the gradual adaptation of HUD (Head-Up-Display) in civil transportation, it was normal for civil pilots or even drivers to use it. Howev-er, any failure warning of airplane HUD was so fatal that required pilot to search for it immediately and make correct response to recover. As the pri-mary indicator of HUD warning, failure flag was well-designed especially in visual coding, which directly affected pilot’s recognition and acquisition of warning information. This research developed ergonomics experiment of HUD simulation interface and designed character size of failure flag as ex-periment variable. The participants were required to perform signal detection task and response to signal (warning occurred) or noise (no warning) accord-ingly. Therefore, both sensitivity and response bias were measured to analyze visual characteristics of failure flag and its influence on HUD warning.The mechanical failures of airplane were mainly caused by unreasonable mecha-nism structures or improper system operations, which was excluded from conse-quence of any manually operation or other outside factors . And the failure flag was shown on HUD when the sensor status was unavailable and the same parameter from different sources did not match . The representative mechanical failure flags were included with indicated airspeed warning, barometrical altitude warning, head-ing warning and ground proximity warning. Moreover, HUD failure warning could be displayed in various channels of visual, auditory and tactile manners, and the first two were most widely used channels . The visual channel of HUD failure warning was normally designed in visual coding technology, which was involved with size, position, salient of blink or textbox frame. And the auditory channel focused on volume and tone of warning voice. Specifically, the minimum character size of failure flag was 1/200 of the visual field distance, i.e., at least 0.18-inch character was required to be shown on HUD at distance of 36-inch away. In addition, the aspect ratio for English letter was required as 3:5 , and its typeface was selected as Arial . To analyze the effect of visual coding on failure flag on HUD, this research firstly developed HUD interface simu-lation, and then carried out ergonomics experiment to validate the influence of char-acter size.3.2Experiment InterfaceAs shown in Fig.3, the experiment interface was displayed on 17-inch computer screen with visual distance of 60cm. The screen resolution was 1440×900 and aver-age luminance was 120cd/m2. The participants were required to interact with exper-iment computer through normal mouse and keyboard. The HUD interface was simu-lated on GL Studio platform, and the experiment was realized using C++ and net communication technology on Visual Studio 2012. Each interaction with experiment computer was recorded and output when the experiment task was completed.3.4Experimental TaskThe participants were required to perform signal detection task during the experi-ment block, followed by the order of Latin Square fashion. And each block lasted for ten minutes with interval of five minutes. Moreover, the failure flag was designed to occur at possibility of 75%, with noise-signal ratio as 1:3. To response each failure flag, they need to press corresponding button of W/A/S/D/E on keyboard according to the specific warning. And SPACE button was expected to be pressed when no warning was shown.ResultsBoth statistics analysis and signal detection method were used to process the exper-iment results. First, descriptive statistics was used to illustrate central tendency and dispersion degree of detection performance. Then, repeated measured ANOVA was used to examine the effect of character size. In addition, sensitivity d’ and response bias were processed to evaluate the influence of failure warning based on signal detection theory.
Keywords: Ergonomics Experiment · HUD Warning · Failure Flag· Character Size · Signal Detection
Cite this paper: