Cooperative System Design and Implementation of Bluetooth and Wi-Fi

Abstract

We are living in a ubiquitous world. People bring a massive number of mobile devices such as smartphones, tablets, watches, wearable bands and they are connected among each other via various wireless communication technologies. Even though some of them are equipped with a cellular interface, BT and Wi-Fi protocols are still the most widely used communication technologies due to their free usage. The co-existence issue of them has been investigated by industries and academia for a long time due to the fact that they operate in the same frequency band. Since BT and Wi-Fi protocols are completely different in terms of the physical and medium access control layer design, previous researches assume that the both protocol stacks cannot understand the signals of each other. We define the approaches based on the assumption as blind-to-the-other type approaches. However, this assumption is not applicable to state-of-the-art handhelds, since they include the both protocols in a device, i.e., co-located. Co-located devices are able to interpret BT and Wi-Fi signals, and thus lead to novel architecture design that the both protocol stacks share the their information via inter protocol stack messages. We define a novel type of approaches based on the assumption that BT and Wi-Fi signals are mutually understandable as aware-of-the other type approaches. With the co-located device, therefore, we can propose better solutions not only for the traditional BT/Wi-Fi co-existence issues, but also for the other issues through the aid of aware-of-the-other approaches. In this dissertation, we design three different protocol designs. First, we tackle the performance of devices equipped with the shared antenna for BT and Wi-Fi protocol stacks. We show that co-location of both protocols provides new opportunity for one protocol to better understand the other and to operate in harmony with the other to avoid mutual interference. We develop an Opportunistic Bluetooth Transmission (OBT) scheme that enables a dual stack device having the integrated module to exploit previously-unused waiting times of the Wi-Fi protocol. We evaluate its performance through not only a model-based analysis, but also a practical implementation of a prototype testbed. Second, we bring a problem of Wi-Fi scanning overhead through measurements. Due to visiting every channel characteristics of 802.11Wi-Fi scanning procedure, time and energy consumption of Wi-Fi scanning procedure are significant. To reduce the scanning overhead, we design SplitScan architecture which makes stations split the scanning channels and share the results among neighboring stations. SplitScan enables stations to select its own scanning channels in a distributed way, and let stations in proximity share their information via BT packets. We compare the performance of SplitScan with 802.11 standard scanning procedure and show our SplitScan significantly reduces Wi-Fi scanning overhead through simulation and implementation. Third, we assume that pre-installed communication infrastructures are damaged and do not operate properly. In this scenario, a Wi-Fi multi-hop network is a feasible solution for delivering streaming traffic owing to its capability of high data rate support without requiring any infrastructure. We tackle that IEEE 802.11 medium access control (MAC) shows poor end-to-end throughput performance due to its use of carrier-sensing multiple access with collision avoidance (CSMA/CA), especially in multi-hop networks. We develop a fully distributed pipelining time division multiple access (TDMA) MAC, termed DP-MAC, that aims to reduce the impact of hidden nodes and unnecessary contentions.We evaluate its performance through extensive ns-3 simulation and show that our DP-MAC outperforms 802.11 MAC in various multi-hop network scenarios.