It’s time for part two in our series of posts focusing on the data that STISIM Drive generates during a simulation drive. We recently discussed the basics of data collection, along with a few simple things to think about when setting up your scenarios.
Today we’ll focus on the data file that’s created, and the various sections contained within that file.
At the end of each simulation run, the program automatically creates a raw data .DAT file. This file contains all the data collected during the simulation. This file will always be created and can vary in size, depending on the data you opted to collect, and the length of the drive. Once the file has been created, you can specify the file name (as a file in the STISIM Drive main console). You also have control over where the file will be stored once created (as a parameter in your configuration file).
A Peek Inside the Data File
In general, there are two types of data that will appear in the data file: metrics, which the computer computes and tracks for you (e.g. number of crashes, time to collision measures, divided attention measures, etc), and raw data (e.g. lane position, vehicles, speed, etc), which you can use to compute your own metrics. It’s your choice, really, as to what kind of data you want to use: pre-computed metrics that are available directly to you, or untamed data that you’ll need to dive into and complete your own calculations.
No matter which you choose, you need to know what exactly is in the data file before you can make an informed decision. For a quick overview, take a look at the data file selections in the list below (as well as some suggestions for why you should care about these particular selections).
General information block that contains the date and time, scenario driven, configuration file used, etc. This is a general record of what was driven and when.
Raw Data Blocks
This is arguably the most important section of the data file because it contains columns of raw data that you can then turn into your own custom metrics. A data file can have multiple raw data blocks that contain different columns of data, depending on what’s happening during the drive.
RMS Data Segments
These blocks simply provide mean and standard deviation values for some of the most common simulator variables (lane position, speed, etc.). By using RMS blocks, you don’t have to use raw data blocks, and the average and standard deviations are computed automatically.
Time to Collision Data
This section provides a list of every vehicle in the simulation drive and gives you the minimum time to collision and range between the simulator vehicle and the driven vehicle.
This is a list of all of the simulated vehicles that were in front of the driven vehicle and how close they got to the lead vehicle.
Divided Attention Data
Reports the response time and corrected response for the simulator’s built-in divided-attention task.
Digital Input Events Data
For systems that accept digital inputs from other devices, the data is listed here. This allows you to track data from external sources, such as other computers.
Vehicle Following Events Data
If a Vehicle Following Event is used, the data collected is contained in this block. You can use the data that is automatically computed, or use the raw data that is generated, to do your own analysis using FFT or other techniques.
Intersection Turning Data
Reports what the driver did at each of the intersections in the scene. One of the intersection parameters is a flag specifying what the driver is supposed to do. Using this data allows you to quickly see if the driver negotiated the intersection correctly.
Polygon Attention Task Data
One of the options in the Polygon Event allows it to be used as a divided attention task. If the Polygon Event is used in this fashion, the data will be shown here. This allows you to perform a divided attention task with your own symbols instead of the simulator’s default symbols.
Turn Signal Usage Data
Contains information on how well the driver was using their turn signals during the drive. However, this data is subjective, so it might be useful – but it also might not.
Internally Triggered Events Data
A complete list of data for each internal trigger (specifically set triggers in the scenario file) that occurred during the simulated drive. This data can be used to determine when dynamic elements begin to do things in the simulator.
Externally Triggered Events Data
A complete list of data for each external trigger (user launched triggers) that occurred during the simulated drive. This data can be used to determine when dynamic elements begin to do things in the simulator.
Eye Tracker Streamed Data
For systems that include Eyeworks eye tracking software, this section includes the data that is obtained from the eye tracker. This section allows you to include eye tracking data in the data file, and can be used to synchronize eye data with driver responses in the roadway environment.
List of Driver Mistakes
Finally, the last section contains a complete list of each of the “mistakes” the driver made and some information about each mistake. This section contains aggregate measures of crashes, vehicle collisions, pedestrian collisions and more. Furthermore, this section allows you to go in and see where every crash, ticket, centerline crossing, etc, occurred.
The Power of Data
When you understand the contents of the data file that gets generated, you’ve taken a huge step in deriving meaning from the simulated drives – and in reducing how much data you need to actually look at. For more information and detailed descriptions of each section of the data file, refer to the Data Manual found in the STISIM Drive Help folder.
Stay tuned! Next time we’ll talk about getting started with turning the data into meaningful metrics!