I have to admit that I am not in this business and use only rationality. Also, I have no data. There is individual reaction delay time, giving perhaps about 1 sec difference across individuals who think the same.

If you see red - @brake applied, but the strength depend on one's speed and distance, ideally to stop just few meters before (for security). If you have low speed and/or rather far away and expect light to change - then slow down gas (or foot off) to arrive not earlier then light will change.

I found some articles on the issue: Drivers' Brake Reaction Times

http://hfs.sagepub.com/content/13/1/23.short

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.372.2748&rep=rep1&type=pdf

In general the reaction time behaves as 0.07 V2 with V is the velocity of the vehicle. 

Dear,

We gather Floating Car Data : in-car registration of location and time stamps of driving vehicles.

We can locally increase the sampling rate to have e.g. a location (longitude, latitude) every one second in the area around a junction.

Penetration rate is a statistical variable but clearly more than 3%.

Be aware that we offer this as a commercial service.

www.be-mobile-international.com

www.be-mobile-international.com

You may find my paper that modeled decelerating patterns of vehicles at signalized intersection based on real driving data obtained by experimental cars of NISSAN Motors. I modeled the decelerating/speeding patterns based on speed and distance from the red signal. I used it to predict preceding vehicle for eco-driving. Here is the main paper used in modeling.

1. M.A.S. Kamal, M. Mukai, J. Murata and T. Kawabe, "Ecological Driving Based on Preceding Vehicle Prediction Using MPC," Proc. of the 18th World Congress of IFAC, pp.3843-3848, Italy 2011.

The model also used in

2. M.A.S. Kamal, M. Mukai, J. Murata and T. Kawabe, “Model Predictive Control of Vehicles on Urban Roads for Improved Fuel Economy,” in the IEEE Transactions on Control Systems Technology, Vol. 21, No.3, pp.831-841, May 2013.

If you cannot download and need copies please let me know.

Best regards,

Urie, you might be interested in two recent papers by Rittger et al. Links to the articles, developed within the framework of the German research project UR:BAN, are attached below.

Best Regards

Ulrich

https://www.google.de/url?sa=t&source=web&rct=j&url=https://www.psychologie.uni-freiburg.de/Members/kiesel/Publications/Rittger_Schmidt_Maag_Kiesel_in_press_CognTechWork.pdf&ved=0CB8QFjACOApqFQoTCPPkvtGp_sgCFeclcgodWaAE9A&usg=AFQjCNHjKmcBSYl-M-CIbPcsDO86LHu7hg

https://www.google.de/url?sa=t&source=web&rct=j&url=http://www.hfes-europe.org/wp-content/uploads/2014/11/Rittger.pdf&ved=0CFQQFjASahUKEwjmpqX9p_7IAhWkwHIKHT2-AEo&usg=AFQjCNGAQXtMkVMHENPhhuePyrxuz94eBw

The dataset we used for the analysis in this paper may be useful, but I don't know how you would be able to get traffic signal data in sync with the GPS logging?

Article Using on-board logging devices to study the long-term impact...

Can show me the research literature on :

+ Impact temperature and humidity to the ability to work the flexible pavement surface .

+ Influence of bitumen to vicious streak flexible pavement surface subsidence wheel .

This question belongs to 3 years ago but I'd still like to add my comments based on my understandings for the prospective interested readers:

1) For the purpose of your research (and possibly any other aspect of human factors in microscopic level), the accuracy of the data in itself is very important. The NGSIM data is also shown to be riddled with systematic errors due to the rotation of camera and the associated image processing methods. Yet, these errors may not affect your research if you are mainly interested in macroscopic state variables or some aspects of car-following behaviour. But if you are trying to extract additional human factors, these data may not be a good source. the following papers presents more discussion on the accuracy of the NGSIM data:

i) On the assessment of vehicle trajectory data accuracy and application to the Next Generation SIMulation (NGSIM) program data

V Punzo, MT Borzacchiello, B Ciuffo

Transportation Research Part C: Emerging Technologies 19 (6), 1243-1262

ii) Trajectory data reconstruction and simulation-based validation against macroscopic traffic patterns

M Montanino, V Punzo

Transportation Research Part B: Methodological 80, 82-106

2) GPS-based Floating car data (FCD) may not again suit what you're looking as the spatial & temporal accuracy of these data may be below what you may need. Extended floating car data (XFCD), however, can compensate for the limitation of FCD as you have information of the leading vehicle. But this is helpful if at least the temporal resolution of the data is above your human reaction time...

3) If you are developing a new methodology, probably the best way (to my knowledge) is to test it on reconstructed trajecoty data by commercial softwares such as VISSIM & AIMSUN to minimise the effect of systematic error in the process of data collection. You should, however, be advised that whatever results you are getting with using these data are based on Viedemann (VISSIM) and Gipps car-following models (AIMSUN) and may not produce adequate insights about the actual human factors.