Development of a personal computer-based secondary task procedure as a surrogate for a driving simulator

This research was conducted to develop and test a personal computer-based study procedure (PCSP) with secondary task loading for use in human factors laboratory experiments in lieu of a driving simulator to test reading time and understanding of traffic control devices such as changeable message sign (CMS) messages. Using MicrosoftRTM Visual C#RTM , a PCSP was developed where subjects were shown CMS messages while simultaneously deactivating randomly displayed buttons in an on-screen control panel which served as a secondary loading task. The subject secondary task workload could be varied by increasing or decreasing the rate the buttons appeared in the control panel. The PCSP was designed to: (1) Display a wide variety of CMS messages including two phase and/or flashing messages, (2) Provide a push-button secondary task that could be varied in button display rate, and (3) Automatically store subject performance data such as reading time and button-pressing time for later retrieval and analysis.; One-hundred-twenty-six subjects were tested within the PCSP and in the Texas Transportation Institute's (TTI's) Driving Environment Simulator. The subjects were subdivided into three subgroups to compare performance between the driving simulator and the PCSP with respect to differences in reading times, comprehension, and preference for alternative CMS messages. The study consisted of comparing reading times, comprehension, and preferences for each of five types of CMS messages with an alternative message. Three different levels of secondary task loading were selected for the PCSP: no secondary task, buttons displayed at a rate of 0.625 button/second, and buttons displayed at 0.909 button/second.; Data analysis revealed that only three instances of statistically significant results were found between the PCSP with no secondary task and the driving simulator out of 41 hypotheses tested. Additionally, only three differences were found for the PSCP with 0.625 button/second secondary task display rate and six for the PCSP at the 0.909 button/second secondary task display rate. The few differences between the three display rates revealed no trends, and indicated that all three of the versions of the PCSP seem acceptable for use in studying CMS messages in order to find comparable results to TTI's driving simulator.