Category EN P19 Linear Analysis of Times Series to Identfy Optimal Time Parameters

for Lane De

Abstract The goal of this project is to determine the time parameters for the two-

second rule first developed by Delphi and implemented by Ford Motor

Company to control the trigger of the adaptive Lane Departure Warning

System (LDW). The two-second rule states that the hypothetical driver

must be in the 'pose not forward' position for two minutes and in a state of

distraction for one minute inside the lane. 'Pose not forward' means that

the driver is not facing forward; this is detected by a camera focusing on

specific facial features, such as a nose or ears. If the time limits are met

by the driver then the LDW system assumes that the driver is exiting the

lane unintentionally and is triggered. However, in Ford VIRTTEX lab's study

of the LDW several of the subjects were not warned about the time or

were warned at the wrong time. Six of the total subjects received a

delayed warning because of the two-second rule. In order to identify the

optimal time parameters for the LDW system several MATLAB scripts will

be coded and utilized. The first script will analyze the variables of each of

the forty subjects to identify the amount of time that the subject was 'pose

not forward', the amount of time that the subject was in a state of

distraction and the time difference between when the LDW was triggered

and when the subject actually exited the lane. The second MATLAB script

will identify the 'noise' of the LDW system. 'Noise' is similar to percent error

of fractional error; it represents the technological faults. While identifying

the optimal time range for the LDW to be triggered might be finding a

specific number, it might not be the perfect trigger rule. Therefore, this

project will also entail a third script that identifies the cumulative sliding time

window. The cumulative rule will combine all of the driver's 'pose not

forward' time to account for technical or behavioral mistakes. While the

goal of this research project is to utilize MATLAB to identify the optimal time

parameters for the LDW trigger, the purpose of this project is to determine

the safest and most accurate method of applying the LDW technology to

everyday life.

Bibliography Ihttp://www.volpe.dot.gov/hf/roadway/saveit/docs/datafus_rep.dochttp://

www.fmcsa.dot.gov/facts-research/research-technology/report/lane-

departure-warning-systems.htm
First Previous Next Last