Study on a method for evaluating the safety of the braking control algorithm for automated driving system when following 2019-01-1015
The purpose of this study is to develop the evaluation method for the safety of the braking control algorism for the automated driving system under the mixed traffic flow consisting of automated driving vehicles and vehicles driven by driver.
We consider that the automated driving vehicle should be controlled like becoming familiar with this mixed traffic.
Therefore, in the safety evaluation of the braking control for the automated driving system in follow-up driving, the influence to the driver which is given by the automated driving vehicle which run ahead of the vehicle driven by driver is an important evaluation index.
First, we analyzed traffic accidents which were occurred in Japan to specify the suitable traffic environment for evaluating the safety of the braking control algorism for the automated driving system in follow-up driving.
As a result, it was founded that this evaluation should be conducted in the situation that the vehicle driven by driver run behind the automated driving vehicle on the road section of uninterrupted flow.
Second, driver's braking operations was measured using actual vehicles in this situation. We developed the method of generating sample algorisms of the braking control based on this driver's braking operation.
Finally, we developed the method of finding out the most suitable range of parameter of braking control algorisms by evaluating these sample algorisms based on results of subject experiments.
This evaluation method uses the driving simulator. The automated driving vehicle installed the sample algorithm of the braking control runs ahead of the vehicle driven by subject in the driving simulator. Subject evaluate the feelings of danger for the braking of the automated driving vehicle.
The most suitable range of parameter of braking control algorisms for the automated driving system in follow-up driving is cleared by classifying results of this subjective evaluation.