U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations
This magazine is an archived publication and may contain dated technical, contact, and link information.
|Publication Number: Date: Winter 1994|
Issue No: Vol. 57 No. 3
Date: Winter 1994
The Federal Highway Administration's (FHWA) Human Factors Laboratory has operated a Highway Driving Simulator (HYSIM) at the Turner-Fairbank Highway Research Center (TFHRC) since the early 1980s. The HYSIM is a fixed-base, interactive driving simulator that uses computer-generated imagery for its visual display. The HYSIM is an excellent test bed for a wide variety of studies of driver performance.
This article discusses the utility of driving simulators for human factors research, HYSIM's evolution over the past 10 years, ongoing and planned HYSIM research efforts, and FHWA's anticipated direction in the field of driving simulation.
Driving simulators offer a number of advantages over other testing environments. Unlike field studies, simulators offer complete control of environmental factors and are highly cost-effective for setup and data collection. Furthermore, simulators offer a safe environment in which to test conditions that may be too dangerous to evaluate in a real-world environment. Unlike other laboratory situations, however, simulators offer a dynamic driving environment with workload and tasks similar to actual driving. Driving simulators complement other laboratory methods and field testing, in many cases providing the best of both worlds.
A wide variety of features and configurations can be applied to driving simulators. Driving simulators can be fully interactive, meaning that all the control actions taken by the driver are reflected in the scene. These systems allow a wide range of driving situations to be evaluated in a realistic setting Partially interactive or part-task simulators, although more limited in the range of experimental situations, are less costly to develop and maintain and can provide useful information for a number of human factors issues. Computer-generated imagery allows tremendous flexibility; any required roadway and surrounding environment can be developed, but the visual image may lack the rich detail seen in the real world Real-world footage, through film or video mediums, can display the full visual complexity of the driving environment, but scenes are limited to existing roadways and are not fully interactive. Motion-based simulators provide increased face validity to the system. They are important for laboratory investigations of emergency responses and vehicle handling. However, compared to fixed-base systems, motion-based simulators require significantly greater fiscal resources, computer power, space, maintenance, and staffing.
In 1980, Systems Technology Incorporated completed a feasibility study for the Department of Transportation (DOT)/FHWA Highway Driving Simulator. Subsequently, HYSIM was built and tested under federal oversight in California. The original system was installed at TFHRC, tested and became operational in 1983. The HYSIM was designed and built in a modular fashion, allowing subsystems to be upgraded as the state of the art in various technologies advanced.
Basically, HYSIM consists of a car cab that provides a realistic driving environment, a variety of computers to control the simulation, and two primary visual systems. A wide-screen projection system is used to display the roadway, the surrounding environment, and other vehicles. Four 3imm slide projectors with zoom lenses and affiliated yaw mirrors add signs to the scenario. The capability of showing high-resolution signs in real time is what continues to set HYSIM apart from other driving simulators.
The original HYSIM consisted of nine modules. The Car Cab Module, still with HYSIM, is a modified 1980 Ford Fairmont. Except for the engine, drive train, and wheels, the car cab is intact. It provides a realistic automobile environment for the driver. Driver manipulation of controls in the car cab is reflected in the simulated driving environment.
The Graphics Computer Module was a DEC PDP 11/34. It received input from the car cab controls to provide updated car position and velocity to the scenario computer for navigational calculations. The graphics computer also controlled the vehicle dynamics of the HYSIM.
The Scenario Computer Module was also a DEC PDP 11/34. This module provided primary control of the experimental scenario, performed navigational calculations, controlled other modules, and executed data collection. The Graphics-Generation Module was an Evans and Sutherland Picture System 2. This module transformed the aerial view of a predefined roadway generated by the graphics computer into a perspective view. The output was displayed on a high-resolution color monitor.
The Roadway Projection Module consisted of a Sharp XCRA color television camera and an Aquastar 80090 television projection system. The camera viewed the high-resolution display of the graphics-generation module and output the image to the Aquastar, which was mounted in a gantry above the car cab. The Aquastar then displayed the roadway onto a screen mounted in front of the car cab.
The Sign-Generation Module consisted of four Mast Random Access slide projectors, also mounted in the gantry above the car cab. Affiliated 7:1 zoom lenses allowed the static 3imm slides of signs to grow in size as the driver approached them, and computercontrolled yaw mirrors maintained the appropriate lateral placement of signs on the roadway. The signgeneration module was monitored and controlled by the scenario computer and projected realistic sign images onto the screen in front of the car cab.
The Sound-Generation Module received output from both computers. It provided crash sounds, sirens, wind noise, engine sounds, and tire squeals to the simulation through speakers located in the doors of the car cab.
The Psychophysiological Module was a Gould system that allowed for collection of various physiological measures, including respiration rate, heart rate, and galvanic skin response.
The Operator Control Center Module was located in a separate room from the vehicle. This workstation included a variety of monitors that allowed experimenters to remotely monitor the progress of experiments and to view the projected image or the subject. All of the electrical wiring for the HYSIM I/O devices was routed through the control center to facilitate troubleshooting and the addition of ancillary devices.
This first-generation HYSIM represented the state of the art in driving simulation in the United States, and although highly sophisticated for its time, it provided a relatively sterile visual environment. The graphics-generation module displayed roadway delineation, and the sign generators provided highway signs. The image was that of a flat road at night with no surrounding environment or other vehicles.
Even with these limitations, the first-generation HYSIM provided a safe, controlled driving environment and was a highly successful test bed for numerous important research investigations.
Many of these studies required increased capabilities in the HYSIM, and over the years, additional modules were added. The Space Module, which provided the capability to simulate active signs and signals, was added in 1985. By rapidly alternating between two slides, this module allows signal lights to change and signs to flash. In 1988, the Vehicle Module was added to provide another vehicle in the scenario. This system used a video mixer to combine the roadway image with an image of a model car or truck for presentation through the roadway projector.
Individual subsystems in the HYSIM were also upgraded over the years. The zoom range of the sign generators was increased; the quality of the roadway projector was improved; the size of the screen changed; and the projection equipment was relocated to provide rear-projection images of better quality and brightness.
Although these changes increased the utility and realism of the HYSIM, the basic system remained unchanged. In the early 1990s, a major system change was made to the HYSIM, resulting in the second-generation HYSIM.
By 1990, the state of the art in computer graphics made it possible to costeffectively simulate complex visual scenes in real time. The graphics-generation module was replaced with a polygon-based system. This system allowed for display of color images for the roadway and the surrounding environment, significantly enhancing the utility of the HYSIM.
The second-generation HYSIM retains many of the features of the original system and uses the same basic modular design. The Computer/Graphics Module is now contained in a single system. The second-generation HYSIM is based on a STAR Graphicon G2000 Image Generator using a Silicon Graphics IRIS 4D/35TG workstation as the controlling host. This system can display 3000 textured, anti-aliased polygons per frame at a 30-Hz frame rate. The workstation provides all control functions for the experimental scenario, performs navigational calculations, and controls peripheral devices and data collection. The vehicle dynamics simulation is housed in this system as well.
The Roadway Display Module is fed by the image-generation system and outputs the image through a BarcoGraphics 1200 video projection system to a screen placed in front of the vehicle. The input signal is a 1280 X 1024 pixel image, which is directly routed to the projector via an RGB (red-green-blue) signal. The projector is located behind the screen and is positioned to present an angular field of view of approximately 70 degrees horizontally and 35 degrees vertically.
The basic design of the Sign-Generation Module has not changed, but the affiliated equipment is significantly more sophisticated than the original These devices are now located behind the rear-projection screen located in front of the car cab. The system now uses four Navitar 750-watt Xenon arc lamp 35-mm randomaccess slide projectors and 17:1 zoom lenses. The affiliated yaw mirrors can move both laterally and vertically to allow signs to rise over a hill.
A small-scale Motion System Module has been added to the system. Four pneumatic pistons, with affiliated coil springs and shock absorbers, are located in each wheel well of the car cab. This system simulates the normal vibrations experienced while driving and can provide minimal car cab pitch for each corner of the car cab.
A Grass Model 12 Neuro-Data Acquisition System has been added to the Psychophysiological Module to allow for conditioning and recording of electroencephalogram data.
Minor changes have been made to the Car Cab Module to enhance the feel of the controls. A high-torque servo-motor and a system of cables and weights have been added to increase the realism of the steering wheel system.
The Sound-Generation Module, the Operator Control Center Module, and the Space Module are essentially unchanged from the original design.
The significant changes between the first- and second-generation HYSIM have greatly enhanced the utility of the system for human factors research. The new graphics system provides the capability for simulation of daytime scenarios that were lacking in the original system. Also, with the capability to produce trees, buildings, and other environmental features, the visual scene is much more representative of real-world driving, and it increases the visual loading on the driver. The new graphics system also permits adding up to three other moving vehicles in the scenario, increasing the vigilance required by the driver, and providing a wider range of experimental situations.
The upgraded sign generators increase the brightness and visual range of signs in the simulation. The additional vertical yaw mirrors will allow vertical curves to be added to HYSIM scenarios.
Other enhancements, including the motion system and the steering wheel torque system, provide increased face validity, and the upgraded psychophysiological system provides a greater range of data availability than the original system.
Second-generation HYSIM validation study
Because the second-generation HYSIM is significantly different from the previous version, another validation study is required To ensure that results obtained in the HYSIM are generalized to the real world, drivers' responses and behaviors in the HYSIM must, within certain design constraints, correspond to their real-world responses.
Subjects will drive a 1imile scenario encompassing various speed limits, road delineations, highway signs, and traffic volumes -- both on the road in TFHRC's field-test vehicle and in the same scenario programmed on the HYSIM. Driver performance measures will be collected in both settings and systematically compared. This will provide a clear understanding of HYSIM's validity as a research test bed and serve as the basis for the design and implementation of future HYSIM enhancements.
Factors affecting vehicle headway selection
The distance a driver travels behind a lead vehicle is referred to as headway. Understanding how external factors influence headway selection has critical safety and efficiency ramifications The amount of headway a driver selects may function as a behavior-based safety surrogate or predictor of rear-end collisions. Furthermore, if external highway features influence headway selection, highway design elements can be modified to assist drivers in choosing appropriate headways for specific situations.
Subjects from a variety of age groups will drive a scenario that encompasses a variety of roadway features. Geometric design, environmental complexity, and behavior of other vehicles will be manipulated to determine which factors affect driver headway selection.
Partial sleep deprivation and drivng performance
Extended periods of partial sleep deprivation appear to affect cognitive skills more than psychomotor skills. It is hypothesized that operators of heavy vehicles may be sleep-deprived on a regular basis, and fatigue is often cited as a contributing or causal factor in truck accidents. Drivers of personal vehicles also operate under conditions of partial sleep deprivation, but controlled investigations of the effect of sleep deprivation on driving performance can be dangerous to perform. FHWA is planning a HYSIM study in collaboration with staff from the Walter Reed Army Institute of Research to investigate the effects of various levels of sleep deprivation on driving performance.
Subjects will drive a simulator scenario containing a number of emergency events under both normal sleep and partially sleepdeprived conditions. They will be allowed one, two, four, or eight hours of sleep per night over the course of several days before participating in the second simulator run. A variety of psychophysiological measures, including brain activity (electroencephalogram) and heart rate will be collected and analyzed. Driving performance measures will also be analyzed, including appropriateness and reaction time for emergency responses, speed, and lateral placement. Results will delineate the effects of various levels of partial sleep deprivation on driving performance.
HYSIM is particularly suited to the needs of FHWA's Human Factors research program. Highway signing research will remain an important component of the program, but the HYSIM will also be used in other areas. Older driver issues will continue to be addressed in the HYSIM, as well as important issues related to the design of Intelligent Vehicle-Highway Systems. Investigations in new human factors program areas, including hazard identification and younger driver issues, will also be performed in the HYSIM.
In the near term, HYSIM upgrades will include the capability of simulating vertical curves and the continued development of additional intelligent, interactive vehicles that behave like real traffic Future longer term enhancements will be based on the results of the validation study and a careful review of research needs in the FHWA human factors programs. When system components become more sophisticated and less costly, they will be added to the HYSIM as research needs dictate.
Over the course of HYSIM's development, a number of other driving research simulators with different capabilities have been developed elsewhere. As needs arise for simulation research that requires higher fidelity or different capabil ities (su ch as wider field of view or motion) than available in HYSIM, the human factors team will pursue other simulators as test beds. For example, the University of Iowa currently operates a motion-based driving simulator, the Iowa Driving Simulator (IDS), which is being used for ongoing FHWA research efforts.
HYSIM, in all its configurations, is an effective research tool for FHWA human factors investigations. Studies conducted in HYSIM range from traffic control device investigations to driver risk perception studies to Intelligent Vehicle-Highway Systems experiments. HYSIM has been upgraded over the years to enhance the capabilities of the system and to meet the needs of changing research programs. Complex human factors issues in the Intelligent Vehicle-Highway Systems program require further simulation experiments, as will older and younger driver investigations and hazard identification studies. HYSIM provides the necessary capabilities to perform these studies.
Human factors research at FHWA is greatly enhanced by HYSIM. The flexibility of the system configuration permits a wide variety of critical investigations of driver performance to be performed under controlled and safe conditions. The modular system configuration permits the modules to be upgraded as the state of the art improves in various areas. Additional modules have been added over the years to ensure that human factors research needs are met The new graphics system significantly increases HYSIM's capabilities, and new system upgrades will be made to ensure that the system continues to meet the needs of the human factors research program.
Elizabeth Alicandri is the manager of the Human Factors Laboratory at the Turner-Fairbank Highway Research Center in McLean, Va She works in the Information and Behavioral Systems Division in the Office of the (FHWA) Associate Administrator for Research and Development.