Int. J. Networking and Virtual Organisations, Vol. 22, No. 1, 2020 17
Copyright © 2020 Inderscience Enterprises Ltd.
A novel machine learning-based attacker detection
system to secure location aided routing in MANETs
R. Suma*
VTU,
Belgaum, Karnataka 590018, India
and
Department of MCA,
SSIT,
Tumkur, Karnataka 572105, India
Email: sumaraviram@gmail.com
*Corresponding author
B.G. Premasudha
Department of MCA,
SIT,
Tumkur, Karnataka 572103, India
Email: bgpremasudha@gmail.com
V. Ravi Ram
VTU,
Belgaum, Karnataka 590018, India
and
Department of MCA, SSIT,
Tumkur, Karnataka 572105, India
Email: raviramv@gmail.com
Abstract: The proposed work deals with the improvisation of the performance
of location-based routing in mobile ad hoc network (MANET). A machine
learning-based attacker detection (MLAD) algorithm that uses multipath
routing is proposed to facilitate efficient routing even in the presence of
attackers. The proposed algorithm adopts the location aided routing (LAR) to
optimise the search process and to reduce the search area for new routes in
MANETs. Learning automata (LA) is implemented to optimise the path
selection and to reduce overhead in the network. Extended identity-base
cryptography (EIBC) is used for efficient key management in providing system
security. The proposed system implements privacy preserving communication
system (PPCS) for maintaining privacy in end-to-end communication. This
method decouples the location information from the node’s identifier and
abstracts the communication happening among nodes. The simulation results of
the proposed method reveal its reliability and strength in securing LAR in
MANETs.
Keywords: mobile ad hoc network; MANET; learning automata; optimisation;
routing; fault tolerance; identity-based cryptography; key management; privacy
preservation; attacker detection; location aided routing.