Hi, you can have a look in the first chapter of my thesis for an introduction on emotion recognition in the context of brain-computer interfaces: doc.utwente.nl/82072/1/thesis_C_Mühl.pdf
We also recently published a special issue and an freely available survey on the topic: http://www.tandfonline.com/toc/tbci20/current#.VBK-c2OM3Sk
The research was not done in a car, but the general issues with physiology-based emotion recognition (in the wild) are the same and should be interesting. I always also like to recommend Fairclough's work on psychophysiological inferences (references in both docs above).
Thank you Chritian Muhl for your answer and suggestion. I would like to know more about your research that would be helping me in doing my phd research.