Research simulator, continued 2

 

Background

The research simulator software of Carnetsoft has been used in numerous experiments. The development started in the early 90-ties at the traffic research center in Groningen (the earliest description of the software in the literature: Wolffelaar, P.C. van & Winsum, W. van. (1992). A new driving simulator including an interactive intelligent traffic environment. In: Proceedings of the 3rd international conference on vehicle navigation & information systems: 499-506. IEEE Service Center, Piscataway NJ.). At that time it was one of the first research simulators with a script language and dynamic intelligent interactive traffic. The software was applied in a number of experiments on driver behaviour, for example:

– Van Winsum, W. & Brouwer, W. (1997). Time headway in car following and operational performance during unexpected braking.Perceptual and Motor Skills, 84, 1247-1257.
– Van Winsum, W. & Heino, A. (1996). Choice of time-headway in car-following and the role of time-to-collision information in braking. Ergonomics, 39(4), 579-592
– Van Winsum, W. & Godthelp, H. (1996). Speed choice and steering behavior in curve driving. Human Factors, 38(3), 434-441.
– Van Winsum, W. (1996). From adaptive control to adaptive driver behaviour. PhD Thesis. Haren., The Netherlands: Traffic Research Centre, University of Groningen.

Around the year 2000 the software was developed further in ST Software. During the period 2000-2011 research simulators were sold to a number of Universities around the world, and the software is used in a large number of experiments. For example, a large number of Human Factors studies can be found in the literature from the university of Groningen that have been done with a research driving simulator that is powered with this software.

Since 2011 the software was developed further in Carnetsoft, focussing mainly on driver training simulators. The rendering system was replaced by a new system that can be easily modified by the user (with python). In the mean time the software was extended with functionality missing in the previous versions. Since 2013 the research simulator software (database generation and experiment preparation) was developed further to improve user friendliness and the research simulator software was added to the product list as a desktop system.

This research driving simulator can be used for research in other fields than traffic psychology as well. Various research questions in the fields of cognitive psychology or cognitive neurosciences can be examined. In addition to the regular driving simulator facilities, it also has methods to measure workload, such as the standard Detection Response Task. It has an interface to PsychoPy as well. This means you can define various psychological tasks (for example the Sternberg task or the Posner task) and run these while driving specific scenarios. The buttons on the steering wheel can be used for response generation and measured from within the PsychoPy task, while the PsychoPy tasks can be controlled via script functions of the simulator. The PsychoPy task can be overlayed over the simulator images on the left, middle or right monitor.

The flexibility of the script language gives researchers unprecedented freedom to design their experiments fast and efficiently.

Video examples of databases and scenarios

In the examples below, only the forward view is displayed instead of the regular left/forward/right views that are rendered on 3 HD monitors in the actual simulator.

Description: Example of driving in a rural virtual environment. In the simulator this is rendered over 3 HD monitors with 210 degrees horizontal field of view, at around 60 images per second, resulting in crisp and smooth images.Description: Exampe of driving in a village database while encountering a large number of intersections with different priority regimes.

Characteristics

Comes with ready to be used content

Saves you a lot of time to create experiments

tickmarkRoad databases: includes 15 databases that are used in the driver training simulator modules. Because the sources are included, there can be extended or modified: highways, rural roads, roundabouts, towns and villages…. All databases have roadsigns of the country you live in.
tickmarkSimulations/scripts: includes the script source code for all driver training simulator lessons that can be used for experiments, or as example script on how to create traffic, interactions with the driver, store data, etc.
tickmarkStandard tasks: to measure workload (Detection Response Task), carfollowing task, tasks to measure brake reaction time, susceptibility to falling asleep while driving, etc.

Flexibility in developing simulations

tickmark
Flexible script language to create your simulations: generate traffic, send pictures to the monitors, create speech messages, store data, etc.
tickmark
Python source code for the rendering system included
tickmark
Tools to configure the system: enable or disable graphics effects such as shadows, or light scattering effects, configure displays and resolutions, set to left hand or right hand drive (and the corresponding traffic rules), etc.
tickmark
Tools to create road databases or modify existing databases. Easy to add new road signs or add new objects (buildings etc.)

Integration with other applications

tickmark
To other computers. The script language allows you send or read data from UDP sockets, or send data via a serial port.
tickmark
PsychoPy. PsychoPy tasks can be run on the same (simulator) computer, and read the buttons on the steering wheel as response buttons, or be controlled from script via interface functions. All you need to do is include a python script file in your PsychoPy task. Communication is fast via shared memory.
tickmark
Devices. Up to 5 different usb devices can be connected for example to connect a different set of pedals, or button box.

Data measurement and analysis

tickmarkContinuous data sampling is easy: define the sample frequency (for example 10Hz), and the types of data to be sampled from a large list of possible variables: time headway, time to collision, time to intersection, brake pedal, lateral position, yaw rate, etc. You can also create new data functions and store these data continuously with a fixed sample frequency.
tickmarkAnalysis tool to analyze or inspect all continuous data after the trials. Also to export the data to fixed column text files (can be read into Excel).
tickmarkProcess data during the simulation. You can process data during the simulation, for example to compute the brake reaction time, or the standard deviation of lateral position on a specified stretch of road, or the minimum TTC during a certain action of the lead vehicle. This is independent of the fixed sample frequency of the continuous data, but is measured at more than 100 Hz. The data are immediately available, as text files, after running a trial.
tickmarkHigher level data. The student assessment system of the driver training simulator program can also be used to store data in the experiments if you use these simulations. This concerns data such as number of speed violations, number of time the subject did not look to left or right while approaching an intersection, failure to use the indicator, etc. These driver errors are linked to predefined driving tasks (speed control, negotiating a roundabout, overtaking, lane changing etc.)

Advanced simulations

tickmarkGraphics: Night driving, particle effects such as rain or snow, glow mapping for lighting effects while driving at night, effects of fog with varying densities, effects of sun blinding, light scattering effects, shadows. Three rendering channels (left, middle, right) running each on a separate processor with HD resolution for each channel. Each channel has a rear view mirror. Also, top view to check position on the road.
tickmarkTraffic generation and scenario generation: Realistic intelligent interactive traffic that applies all traffic rules. Can be overriden by script. All traffic and scenario generation is written in fast C++ code.
tickmarkAdvanced vehicle model: An advanced engine model and vehice dynamics model applies all relevant forces excerted on the vehicle. This results in realistic force feedback on the steering wheel. Vehicle model data can be modified in script, and output of the vehicle model can be sent to a motion platform.
tickmark3D audio: Realistic 3D audio with doppler effects of approaching and passing vehicles. Sound effects of engine, wind noise, sound of tires, etc.

—Return to previous page—

—For more information, continue on next research simulator page

Go to:

Research simulator- Page 1

Research simulator- Page 2

Research simulator- Page 3

Research simulator- Page 4