中文摘要：

针对无线竞争共享型信道中恶意节点（或自私节点）不公平使用信道资源的退避访问攻击行为，提出一种新的中心极限定理检测算法。该算法能够有效降低计算复杂度，按需调节检测速度和检测精度、并且易于实现分布式联合检测。该文还基于CLT算法建Markov链多步检测模型，进一步提升对恶意节点的检测精度。仿真试验表明，CLT检测算法具有较低的误判率和较快的检测速度，对介质访问控制（MAC）子层退避攻击行为的检测性能明显优于经典DOMINO算法。

英文摘要：

The competitive and shared wireless channel is vulnerable by random backoff attacks from malicious nodes （or selfish nodes）. As a consequence, the channel resource can be occupied unfairly, or the private benefit can be enhanced selfishly. This paper proposes a novel central limit theorem （CLT） based algorithm to detect medium access control （MAC） misbehaviors of malicious nodes. This algorithm is able to decrease the computational complexity, adjust the efficiency and the accuracy of detection on-demand, and implement the collaborated detection in a distributed manner. In order to improve the detection correctness of malicious nodes, the CLT based algorithm can be further enhanced by the Markov chain detection model. Compared with the traditional detection methods, the CLT based algorithm has the characteristics with lower misdiagnose probability and more detection efficiency. Simulation experiments show that the CLT based algorithm is superior to the DOMINO algorithm for diagnosing backoff attacks misbehaviors, in terms of detection correctness and convergence speed.