Contributions à la détection des comportements malhonnêtes dans les réseaux ad hoc AODV par analyse de la confiance implicite

An ad hoc network is a collection of mobile devices that communicate in a self organized wayusing wireless network interfaces without neither centralized administration nor fixed infrastruc-ture. In such a network, nodes must cooperate with each other so as to extend their transmissionrange to reach distant nodes. This cooperation requires a specific ad hoc routing protocol to es-tablish and maintain routes between nodes. Ad hoc routing protocols are based on mutual trustbetween collaborating nodes and suppose a correct behavior. However, as some nodes that legiti-mately belong to the network may lie so as to manipulate it to their advantage, every node shouldconsider its environment as hostile and implement accordingly its own security mechanisms toprotect itself against internal (i.e. a priori trusted but in fact not trustworthy) dishonest nodes.In this thesis, we are interested in detecting misbehaving nodes within the ad hoc routingprotocol AODV. We propose and implement a detection system based on implicit trust relations : anode implementing this system collects its neighbors’ routing messages and reasons about them todecide on their trustworthiness. We also evaluate our implementation, and, based on simulations,show that the system we have developed to detect dishonest behavior is efficient. We also studycases of false positive and we propose a solution to limit their occurrences.