Exploiting PHY Layer Information to Combat Cross Technology Interference in Low-power Wireless Networks

The proliferation of a wide range of wireless devices operating in the ISM bands is becoming a major challenge for emerging low-power wireless networks in indoor applications. Interference can significantly reduce the overall delivery ratio of such networks, affect their availability and drain their resources. In low-power wireless networks, potential interference can be divided into two broad categories, intratechnology and Cross-Technology Interference (CTI). The latter interference has become a core concern for low-power wireless networks. Frequently used interference mitigation techniques are inapplicable or inefficient for low-power single antenna wireless devices in the presence of CTI. Hence, novel mitigation schemes which can help low-power wireless devices combat CTI are needed. This thesis presents a CTI aware mitigation mechanism, a cross-layer design that harnesses the PHY layer information to help 802.15.4 devices mitigate interference. Besides decoded symbols, physical layer information, such as signal power, soft demodulated bits and number of unmatched bits of decoding, is delivered to upper layer. Symbol error estimation approach which jointly utilizes signal power and unmatched bits is proposed. Based on the symbol error estimation results, symbol errors are categorized into different error patterns with extracted characteristics. Proper interference mitigation techniques, such as bit level interleaving, forward error correction coding, packet merging and diversity combining, with adaptively assigned parameters are chosen according to error patterns and characteristics. Evaluation results show that our symbol error estimation approach has a stable performance with both precision and recall about 80% and our CTI aware mitigation schemes can successfully decrease packet error rate by about 40% for wireless camera interferer and 25% for microwave oven interferer.