SAFH - Smooth Adaptive Frequency Hopping

Abstract

Wireless systems based on WLAN (802.11), ZigBee (802.15.4) and Bluetooth (802.15.1) are continuously deployed in new applications covering consumer, industry or medical fields. Especially, Bluetooth is recommended by the Health-Care-Organization for medical applications as frequency hopping is considered as a robust scheme. However dealing with frequency-dynamic sources of interference in the 2.4GHz ISM band is important due to the increase of wireless devices. Adaptive frequency hopping (AFH) suggested by the Bluetooth standard and implemented in many of todays products identifies and avoids using bad channels. It is a good and established coexistence mechanism in the presence of frequency-static sources of interference such as WLANs when the 2.4GHz band is not crowded. However, AFH is facing problems in a crowded 2.4GHz band, especially when the interference is dynamic. We developed a cross-layer algorithm SAFH (Smooth Adaptive Frequency Hopping) that is inspired by entropy maximization and the conventional Bluetooth AFH. SAFH assigns usage probabilities to all channels based on an exponential smoothing filter for frame error rates to estimate and predict the channel conditions. The application layer can adapt SAFH by parameter settings in a cross-layer approach. SAFH achieves low average frame error rate and responds fast to changing channel conditions if required from the application. Simulative Evaluation in the presence of different types of interference (802.11b, 802.15.4 and 802.15.1) shows that our algorithm outperforms conventional frequency hopping and AFH. Additionally, SAFH works smoothly and stable exploiting frequency diversity compared to previous approaches like entropy-maximization based adaptive frequency hopping and Utility Based Adaptive Frequency Hopping (UBAFH).