by Alapan Chaudhuri
Due to the huge surge in population and urbanisation, the use of transportation has seen a voluminous rise and as a result, large number of problems intertwined with it have also surfaced. One of such major problems is accidents - many of which occur due to drivers' errors.
The absence of proper counter measures against accidents result in huge loses of life and economy. However, with the advent of Internet of Things and through it the evolution of Intelligent Transport Systems, we now have a way to tackle this problem.
The paper , on which this report is based, aims at attempting to solve this problem by monitoring the driver's behaviour and immediate surroundings influencing it.
In order to keep a track of all different factors, different sensors are needed and in order to make correct decisions and predictions proper analysis of the heterogeneous data in real time is necessary.
As mentioned in the paper, Fog Computing  provides an interesting solution in this case. It can reduce latency to a great extent and help in making processing and generating of data much faster and closer to real time, by virtue of its positional advantage.
The main drawback is the absence of a responsive mechanism that is both real-time and effective .
In the proposed layered architecture, the layers are Perceptions Layer, Communication and Middleware Layer, Application Layer and Stakeholders.
Functionality: Data Acquisition
In the perception layer, the following sensors are present to collect all relevant data:
Functionality: Communication, Analysis, Event Detection and Trigger as well as Warning and Reporting
Fog upon gathering data from the Perception Layer and runs road safety algorithms to determine if an alert should be generated based on certain well defined thresholds. The entire responsibility of interpreting and analysing data lies on fog.
Unnecessary data communication that might originate due to the heterogeneity of information obtained from the different sensors used is also handled by the fog.
Functionality: Data Storage and Service Creation
The functionality of Application Layer is achieved through the following:
In this section, results regarding the advantage of fog over a cloud-only scenario are presented, based on a simulator setup developed using DEVS .
DEVS is a message-based specification based in C++ with a two-layered structure allowing communication between models through messages. It consists of six functions: constructor for defining I/O ports, initialisation function, external function which is called on receiving a new message, output function for sending out required data through O/P port and destructor which is called at the end of the process to quit all allocations and unused resources. The above also defines the sequence of execution of any DEVS model.
Information about all the models being used and their associated arguments are present in a .ma (Maya) file that helps in keeping the models connected and also provides for the initialisation functions for each model. These models can be thought as classes whose different instances like, user, cloud, fog and broker are used in the simulation.
The statistics present in the paper , clearly show that road accidents, even though they are a reason behind a very significant number of deaths all around the world, can be easily prevented with the help of advancement of technology in the form of Intelligent Transport Systems.
The Architecture presented in the paper - which creates a very efficient system for Driver Behaviour Monitoring - attempts to prevent road crashes, that occur due to driver's behaviour, effectively with the help Fog Computing. The presented results can be very helpful for further studies in this topic, especially in determining the optimum number of fogs and users per fog required to form a sophisticated ITS model based on the provided architecture.
 "Fog Assisted Driver Behaviour Monitoring for Intelligent Transport System" - Mohammad Aazam, Xavier Fernando (Dept. of Electrical and Computer Engineering, Ryerson University, Toronto, Canada)
 "Fog Computing and its Ecosystem" - Ramin Elahi (UC Santa Cruz, Silicon Valley)
 "Using DEVS for Modelling and Simulating a Fog Computing Environment" - Etemad, Mohammad, Md. Aazam and Marc St-Hilaire