An Artificial Immune System for Misbehavior Detection in Mobile Ad Hoc Networks with both Innate, Adaptive Subsystems and with Danger Signal

The successful operation of a mobile ad hoc network depends on cooperation of the nodes in providing services to each other. Nodes act both as terminals and information relays, and participate in a common routing protocol, such as Dynamic Source Routing (DSR) [13]. The network is vulnerable due to faulty or malicious nodes. Misbehavior detection systems aim at removing this vulnerability [1], [2], [3], [4], [6], [7]. Our approach for misbehavior detection in DSR is to use an Artificial Immune System (AIS) [14], [15]. The system is inspired by the natural immune system of vertebrates [10]. The main task of the natural immune system (the IS) is to protect the human body against microorganism invaders and some malfunctioning own cells, while being tolerant to normal own cells, self cells. To accomplish this task, the IS has developed some detection and reaction mechanisms and procedures, which may be useful for solving analogous problems in building an AIS. The work presented here is a continuation of our previous work [6], [7]. In the previous work we proposed a solution for mapping some basic parts of the IS to our AIS: representation, matching, and negative and clonal selection. We implemented and validated the solution in the Glomosim simulator [11]. The system had a separate preliminary phase for collecting selfbehavior examples. This phase had to be run in a protected environment, when there is no misbehavior of the nodes. It is very hard to provide such conditions in a real network. In this work we give three main improvements for our AIS. First, we propose a solution that doesn’t require a preliminary learning phase in the protected environment (the environment without misbehavior). The solution uses analogy with the IS danger signal [8], [9]. Second, we add the innate part of the AIS, which provides fast detection of misbehavior