Page No: 39-54
Abstract : Create a novel network model for mobile ad hoc network (MANET) nodes and actors in
wireless sensor networks to collaborate on event processing. There are two stages in the
development of distributed algorithms: setup and negotiation. The first uses weighted proportional
max-min fairness to initially allocate MANET nodes across event zones, whereas the latter uses a
market-based method to re-distribute the number of MANET nodes based on existing and new
events. A detection technique for malicious packet dropping attacks in MANETs. The mechanism
of the suggested approach is Collaborative Convolutional Neural Network (CCNN), which is
based on the reputation value computed for that node by its neighbours. A node's reputation value
is determined by its network packet forwarding behaviour. The reputation information is collected,
saved, and transferred between nodes before being calculated under various scenarios. A network
simulator was used to test the proposed protocol. The simulation results demonstrate the
effectiveness of its performance. Even in the presence of cryptographic procedures, our method
incurs negligible network bandwidth and latency costs. Moreover, we demonstrate that the
protection is still effective in the presence of misbehaving nodes and routing changes caused by
mobility. While further research is needed to thoroughly evaluate our method, we feel that the
concept of collaborative security in MANETs is a potential future area.
Keyword MANET, TCP/IP, CCNN, ADHOC