Abstract:
Smart technologies have seeped into every aspect of human communication. The increase in population
has escalated the numbers of vehicles on road as well as probability of collision among
them. The development of intelligent vehicles for improving driver experience have introduced
the Mobile Adhoc Technologies for transport systems in the form of VANETs or more recently
referred to as the Inter Vehicular Communication Network. This is an inherently wireless system
which brings its own set of security challenges dependent on its lack of infrastructure, short connection
times and high mobility. VANETs aim to provide user with optimum driving experience
as well as safety on road. The most significant service for the vehicular networks is the availability
of ubiquitous information to the legitimate users because a delay in this life-critical and
time-sensitive network can be fatal in some scenarios. Hence, Denial of Service is the most imminent
security threat for this system. Although, some research has been done for the mitigation
of this issue, data mining approach towards this aspect is minimum. Furthermore, the available
schemes are deficient in one aspect or the other to provide impeccable detection.
The capability of decision trees to timely identify behavioral changes in traffic with low error
rate makes it a powerful detection scheme. Very fast decision tree is a data mining mechanism
which can handle high speed streaming data, suitable for VANETs. The ultimate aim of this
research is to explore a resilient diagnostic technique to mitigate the denial of service attack in
VANETs. The technique will be computationally efficient and error free, to provide a secure environment
for optimum delivery of VANETs applications. Appropriate simulation methodology
for VANETs is imperative for acurate reproduction of attack on the vehicular adhoc system. Real
time mobility simulator and traffic generator have been employed for this purpose. The research
aims to carry out a performance evaluation of the proposed detection scheme through simulations
to test its competence for high speed, sensor data in vehicular adhoc systems and benchmark its
dexterity against other detection schemes. Furthermore, it provides quantitative comparison of
various types of detection schemes. The comparative analysis confirms the ascendancy of the
applied technique with respect to the earlier research in this field.
