Artificial immune system based approach to fault diagnosis for wireless sensor networks

Abstract

Fault diagnosis has been recognized as one of the key issues in wireless sensor networks. Considering distribution feature of sensor node, however, the fault happened in wireless sensor networks is usually random and unpredictable. The conventional diagnosis approaches become increasingly difficult to deal with. As a result, the application is limited seriously. To solve the problem, a new approach based on artificial immune system for fault diagnosis is proposed. The normal and abnormal character patterns generated by a network simulator for wireless sensor networks, respectively, are regarded as the self and antigen of artificial immune system. According to a real-valued negative selection algorithm, the detectors are generated to improve the covering ability of non-self space. Taking detector as antibody, an immunity calculation is executed by the distribution zones of antibody and evolution learning mechanism of artificial immune system. The type of antigen is decided based on the clustering distribution of cloned and matured antibody. The example shows that the approach has better accuracy and the capability of self-adaptive for the fault diagnosis in wireless sensor networks.