As the dog days of summer slowly come to a close, it’s time to stop slacking and get back to the hard work of working hard.
In the simulation business, working hard generally means trying to make sense out of all the data you’ve collected with your simulator. And, depending on your protocol, that can be quite the daunting task!
One of the best things about the STISIM Drive simulator is that it collects a diverse set of data, which means you have a lot of information to pull from. But, the most important thing to understand when it comes to the collected data, is that most of it is raw data. This raw data tells you the state of the simulation at any point in time (e.g. steering angle, blinker state, position of the wheel on the roadway, etc).
The software itself, however, does not compute any performance metrics – it’s up to you to decide what those states mean, or what it means if they change over time.
So this begs the question: How do you deal with the raw data? Try these basic steps:
- Define the metrics that you want to compute (e.g. reaction times, time to collision, average speed and lane position, etc). This will also help you determine the “critical events” and driving sequences that will be incorporated into your scenario.
- Decide on the data that is needed to compute your final metrics (e.g. elapsed time, vehicle speed, vehicle position in the roadway, etc). *Important! Make sure that the simulator can record the data you want!
- Define how the metric will be computed from the raw data (e.g. for average speed, add all the speed data and divide by the number of data points used).
- During your data processing, use some type of a scripting language or tool (we like using Excel macros) to process the raw data and hone it down to the desired metrics, or to plot and look at how different variables change over time.
- Export the data metrics to a high-powered statistical analysis package (SPSS or equivalent) to perform your statistical analysis across drivers and cohorts.
Let’s take as an example the computation of the metrics “average lane position” and “standard deviation of lane position.” In your simulation scenario, you would instruct the software to collect the vehicle’s lane position at each simulation step (a.k.a. the raw data). After the simulation has finished, and you have the data file, you can use this raw data to compute the two desired metrics using standard statistical methods.
There you have it! The basic steps involved in data reduction. Next time we’ll investigate the structure of the STISIM Drive data files and the types of data that are collected. See you then!