Abstract:
Wireless local area networks are used in domestic, commercial and military application. Among different wireless protocols, IEEE 802.11b is the most prominent one. Due to broadcast nature, wireless networks are vulnerable to malicious attacks. Up to certain extent, robust transmission is achieved using channel coding and forward error correction schemes employed by protocol in presence of noise and interference. However, detection of attacks on IEEE 802.11b wireless networks is necessary due to their severe impact on the network performance. Among various malicious attacks, jamming attacks are the most prominent one. The attacks on the radio signal could be protocol aware or protocol independent. Since jamming attacks drastically affect the performance of wireless networks, an effective mechanism is required to cope up with them. We investigate a multi-modal scheme to detect several jamming attacks. This proposed scheme is based on generating profile under different jamming attacks during training session. Our proposed model generates jamming profiles. The profiles generated in detection model are based on packet delivery ratio, signal strength variation and pulse width of the received signal. The model works for both protocol-aware and protocol unaware jammers. We have attempted the proposed model in several jamming scenarios. The achieved results demonstrate a significant improvement in the detection ratio.
