Social engineering and phishing attacks in IoT

I need a research proposal in Social Engineering in phishing attacks in IoT devices (I have to find a specific
smart device to do my research on it).
I have to get the formulation of a clear problem and identify the set of problems with a clear picture in social
engineering and phishing attack in IoT.
please consider these notes:
explain why this problem is challenging? and which part is challenging?
what is the exact prop formulation want to take off to solve these problems? , could be mathematics or
something else (should be three to five exact formulation to discuss).
and this problem has to use actual sensor data; for example, we can do security testing against, actual data set
compared to the simulated data set.
what tools can I use to solves the problem? What is the methodology?
the main contribution is something practical in machine learning (something in python, for example) and adds
three to four sub contributions.
the references should be new, no later than 2017, and some of them should be a survey. and these resources
should be journal or conference papers only.